It is a matter of great privilege for me to be here this morning and at the outset I would like to congratulate both Bell and Globals for organizing this event which is quite appropriate part of the Global AI Summit. Last year, I remember when Prime Minister, our Prime Minister was in the third AI Global Summit in Paris, he underscored the transformative impact of AI on society, emphasizing both its rapid evolution and widespread adoption. He highlighted AI expanding all in shaping politics, economics, national security, and governance.

As we all know, artificial intelligence is no longer a technological advancement; it is a potential. It has the potential to drive innovation, enhance productivity, and act as a force multiplier for security in the defense. It is the strategic capability shaping economic power, geopolitical influence, and national resilience. In fact, AI is becoming the defining capability of the 21st century, reshaping competitiveness and sovereignty. And the decisions which we take today on AI investment and governance will go a long way in defining the trajectory of nations for decades. And that is why India has taken some early lead. And it was not long ago. Let me take a step back that we started talking about AI. It was in late 21 or early 22 when we said there is something called AI and we should start doing something about it. And that is when the first AI platform India was set up. We started the first National Program on AI in 2022 as part of Digital India initiative. And that is when when we launched the India Stack in 2022. This Digital India initiative actually helped India in strengthening its AI policy architecture and digital governance framework. There was a push towards trusted digital public infrastructure and indigenous technology ecosystem. It gained momentum. We developed institutional framework for structured AI adoption across sectors, AI in health, education, agriculture. The Center of Excellence system started announcing the budget and we started working on that. Develop national greater emphasis placed on data governance, DPI, innovative ecosystem and public-private partnership. And this has all, this has led to a strong base for India's AI mission which was launched a year later after we actually held the Global Partnership Summit in December 2023. A collaborative ecosystem between government, startup, academy and industry was set up. A framework for safe, trusted, secure and accountable digital transformation was set up. The foundations built during this period continue to support India's expanding AI ecosystem. Today, while developing the policy framework, we had no doubt in our mind that only a sovereign AI would help India in achieving a safe, trusted, secure and responsible AI. As we all know, rented models will not only have biases but data quality, security and Indian context will be missing in those. India's sovereign AI framework will develop, control and regulate AI systems aligned with our national interest. And it is not isolation; it is strategic autonomy combined with trusted collaboration. Today, India is witnessing AI-driven transformation across sectors, whether it is legal, finance, healthcare, agriculture, governance, manufacturing. You talk about it. While I am now in my new role picking up top executives for the CPSC in India, I am finding that CPs are developing new AI applications which are not only increasing efficiency in operations but they are creating higher value addition across the value chain from incremental to manifold jumps. As startups are disrupting traditional sectors, large enterprises are embedding AI operations. As we know, our policy framework, its, rests on seven important pillars. But let me come back to the AI in defense and strategic. India recognizes the strategic importance of AI in defense. It recognizes the need to leverage AI to enhance combat capabilities, streamline decision-making processes and optimize resource deployment. And that's essential for all our defense forces. In this context, AI offers immense force multiplier potential, enabling superior situational awareness, autonomous systems and data-driven strategies. As technology world must be aware of the latest what is happening in the to maintain our operational superiority in defense and national security context. India's AI policy framework must ensure control over critical data sets, reduce dependence on external AI systems, alignment with national doctrine and values, and protection of our strategic infrastructure. Now, AI is embedded across defense systems which Lieutenant General has talked about. But AI in defense must be controlled, auditable and accountable. It can create havoc if it is not properly controlled. Responsible AI in defense requires human oversight, ethical safeguards, risk assessment framework and clear accountability mechanisms. So, India must pursue sovereign AI with safety, trust and accountability, strategic capability with ethical oversight and indigenous innovation with global cooperation. India has the opportunity to become a leader in AI. We can lead responsibly, strategically and confidently. Thank you. Jai Hind. Thank you, sir. We now welcome one of India's most accomplished and respected defense scientists. She is an inspiration to many, especially women aspiring to lead in science, technology, and national security. Featured by India Today as a leading women defense leader, she has made outstanding contributions to India's strategic capabilities. It is none other than Dr. Chandrika Kaushik, Distinguished Scientist and Director General, Production Coordination and Services Interaction, DRDO. Let's welcome Chandrika Kaushik, ma'am, please. Thank you. Lieutenant General Rana, CNC, SFC, Mr. Alkesh Kumar Sharma, former Secretary, MeitY, Mr. Subhash Gopinath, ladies and gentlemen. First and foremost, I must compliment the Bell and Global team for organizing this very, very relevant topic, at least from the security agencies' perspective, on AI for strategic defense as a part of the AI Impact Summit 2026, which is absolutely imperative today, looking at the way things are changing.

Operational environments for soldiers are progressively becoming complex with incorporation of advanced technologies into military operations. Be it network-centric warfare, multi-domain operations, unmanned combat systems. These complexities are resulting in increased information overloads, reduced reaction times, and quick decision cycles for the human operator. In order to cater to these challenges, it is the need of the hour to introduce autonomy into our military systems to augment the capabilities of the soldier. In June 2018, the Government of India published the National Strategy for AI, focusing on how India can leverage the transformative AI technologies to ensure inclusive growth in all sectors of the society, from sensing to comprehension to action. Today, AI has pervaded numerous end-user applications ranging from computer vision to audio processing, natural language processing, knowledge representation, machine learning, expert systems, etcetera. The AI is being classified into three variants: descriptive AI, which describes what has happened; predictive, which is anticipating what is going to happen; and prescriptive AI, which recommends what to do. With the emerging technological capabilities, AI is geared up to serve as a promising tool to provide enhanced situational awareness to the commanders and operators in highly contested battlefield environments, making our military systems more accurate and our military operations more swift and precise. If we look at the various domains where AI will get integrated, and some of these areas are: ecosystem is working on underwater surveillance, navigation controls, advanced material characterization, hybrid high-fidelity models in the naval systems domain, trajectory and maneuver planning, target identification, threat classification, aerodynamics and controls. So, I have listed out. I think for paucity of time, I will not be able to get into the details. But if you look at it, you look at arms, combat engineering, missiles, strategic systems, electronics and communication systems, computational cyber systems, everywhere there is an application where AI is getting seamlessly integrated. In the near future, AI will find applications in multiple operational areas and some of the use cases, I am just going to deliberate in the next few minutes. With the upcoming challenges of long-range interceptions, reduced reaction times, and saturation of airspace, AI-based decision support systems can greatly augment faster, adaptive, and intelligent decision-making across early surveillance, automated threat classification, quick and autonomous interception, and real-time battle damage assessment. Similarly, in the counter-unmanned aerial systems grid, it can be used to neutralize threats from sub-conventional aerial platforms like drones, loitering munitions, swarm drones, with the network system of AI-powered multi-layer detection and identification, dynamic resource allocation, tracking and engagement of targets through hard kill and soft kill measures. C4ISR operations with man-machine teaming in urban environments can be carried out using coordinated combat operations in complex high-density environments where in challenges of limited visibility, civilian presence, multi-story buildings, and ambush risks. All these can be handled by using AI-enabled advanced surveillance and integrated systems for enhancing the soldier's effectiveness, leading to very, very effective man-machine teaming. With increasing physical, psychological, and cognitive loads in highly contested environments, the future soldier and his equipment will need to be integrated with all nearby sensors and equipment. In such a scenario, AI will help in integrating the sensory information along with the commander's decision-making process through shared mental models. AI will also help in modernizing and coordinating the integrated India's EW capabilities across services with indigenous network and real-time AI-powered electronic warfare suits like smart jammers, cognitive radios, and cyber-secure infrastructure. Weight and size optimization for performance enhancement and operations in extreme environments call for RAD or new-age materials where AI-based modeling and simulations will help in synthesizing these materials, providing higher accuracy, help with adaptive manufacturing, predictive quality control for high-fidelity models, and so and so forth. In defense, reliability and robustness is a critical system requirement to prevent mission failure and unintended consequences. Ensuring that AI applications are safe, resilient to attacks, and aligned with the ethical values is imperative for our defense forces. In order to address this, DRDO has identified five core trustworthy AI principles: reliability and robustness, transparency, fairness, privacy, safety and security, which address various vulnerabilities experienced by AI systems during different life cycle stages of development. The trustworthiness principles have been further augmented with 11 qualifying criteria, and the implementation process has been duly published as a set of two policy documents, namely, Evaluating Trustworthiness in AI Systems: The Framework Document and The Guidelines Document. As we know that AI is as good as it is trained. It heavily depends on training data, most of which is taken from the open source. With respect to India, with its varied geography, essentially needs domestic military data infrastructure for training, validation, and testing data sets. Imported V&V tools may have vulnerabilities. Also, the defense models will be classified and sensitive and need to be validated in indigenous environment. And indigenous AI verification and validation framework will ensure that all data remains within our sovereign boundaries and Indian control, adhering to domestic policies and regulations. Indian military leadership will need interpretable AI decisions with explainability, human trust factors, and ethical considerations in local operating scenarios. The domestic framework is aligned with existing Indian policies and standards published by DGQA, SEMILC, BIS. All these will be required to be aligned. This framework has been officially released last month and is available for the relevant stakeholders and is geared towards providing a unified approach to bringing resilience and trustworthiness in AI-powered military systems. Moving forward, we need a whole-of-nation approach for faster progression and utilization of AI-enabled solutions in defense domain. It is our expectation that academia will establish validation and verification maturity models for each AI and explainability. And we expect industries to play a leading role in developing indigenous AI V&V toolkits, toolchains, including model probes, adversarial testbeds, etcetera. DRDO can support in building standardized defense AI test suites for each domain and use case scenarios. And the quality assurance agencies need to instrument, need to formalize the AI V&V test labs and sandboxes for qualification and airworthiness. Armed forces may need to institutionalize AI safety requirements as a part of the acquisition life cycle process. And Directorate of Standardization is expected to define and enforce the minimum mandatory V&V protocols for operational scenarios. We expect that NIC will be required to build infrastructure for defense-grade secure data repositories for training, validation, and testing. The roadmap towards self-reliance in defense AI lies in establishing a national AI defense strategy. And General Rana spoke about it, detailing various aspects of autonomy, decision support, threat analysis, creating dedicated structures for integrating all the stakeholders in the ecosystem, using the ethical guidelines for ST, and the domestic data infrastructure will need to be ramped up to ensure secure and federated data spanning ISR sensors, battlefield, and historical combat data. We will also need to work on labeling and annotation standards along with standardized processes which will need to be adopted across platforms and operational scenarios. A clear policy on sovereign data should be formulated and implemented at the earliest to prevent any dependency on foreign data and to ensure data residency, specifically for the defense domain. The model design must cater to explainability, adversarial robustness, and fail-safe mechanisms. And dedicated indigenous protocols for domestic combat scenarios should be adopted for verification and validation. I will just finish in another half minute. Low latency AI interface models, hardware accelerators, and trusted environments will be required for deployment and as enablement for resilience. Cyber hardening will be required for ensuring resilience to spoofing, jamming, and cyber attacks. It is proposed that adaptive combat command support may be adopted where in AI must augment and not replace human decision-making in mission-critical tasks. And a redundant human-in-the-loop conventional override systems be included for backup and failover systems. Together, I am sure we can use AI very, very effectively and efficiently in the defense domain for building strategic capabilities. Thank you. Jai Hind. Thank you, ma'am. That was truly inspiring. Today's keynotes from all three dignitaries have been extremely insightful. My sincere thanks to Lieutenant General Dinesh Singh Rana, Shri Alkesh Kumar Sharma, and Dr. Chandrika Kaushik for sharing valuable perspectives. May play also a role in strategic deterrence. As we are talking about nuclear deterrence, we have seen in Cold War the mutual assured destruction. Now there is a talk about the mutual assured AI malfunction, which is already been talked about. And this may be actually a reality. ASI, that is a super intelligence, having that kind of tremendous power, could also be used by a country which possesses this capability against a weaker country in form of what we have seen in nuclear domain, nuclear brinkmanship on nuclear blackmail. Now, when we talk about non-proliferation of these capabilities, in the nuclear domain, it is possible, presently not very successful, but we could have some non-proliferation worldwide, so that non-state actors do not acquire them. But the same would be quite difficult because of obvious reasons when we talk about super intelligence, once it comes into being. Now, when we talk about this artificial intelligence use in strategic domain and national security, there is a lot of usage of the same. In my present appointment as Director General DIA, I have been exposed to many of open-source intelligence systems. If they combine well with the sensors of all the three services, cyber and space, and also some of the other elements like what NASA has developed, the artificial light, the light analysis software, you can really identify the hotspots emerging in your neighborhood or in your own area. So that kind of thing which is already in place in certain places can be further developed by us in times to come. Now, in cyber, when we talk about AI, inherently it is more offensive in nature. In cyber, we have seen the GTG 102 attack which happened in September last year, the autonomous cyber attacks by Chinese state actors against the US systems. But at the same time, the cyber capabilities can be very effectively used to identify the vulnerabilities and plug the gaps. Now, when we talk about other usages, I think you are quite aware, right from the basic usage of logistics and preventing and maintaining your systems from the AI software. We can, we are trans-command and control systems, improving our maritime domain awareness at tactical levels, electronic warfare jamming, very effectively we can do that. And what we have seen in Operation Sindhu, the AI can now be very effective in combating misinformation which is in vast numbers. We saw in Operation Sindhu. Now, the question is, what we need to do? We need to do. And as India, when we talk about the capabilities. First of all, I would suggest that we should have, we should develop an AI security doctrine. And we should use our resources where actually, as per doctrine, where it is required more. And second is, if we take example from Russia-Ukraine conflict. Russia, there are a lot of systems which are getting deployed without waiting from top-down capabilities, battlefield low-cost AI-based systems which are easily scalable and are cheap are getting giving good results. They are increasing the kill chains of these autonomous systems and drones. So, this need to be deployed, while we work on the higher systems of artificial intelligence. Now, taking some example back, when I was commanding a corps in Four Corps, in Arunachal, where I was responsible for security of Line of Actual Control of Tawang and area around. I remember we had a very, we had developed a small artificial intelligence software, low-cost. But when there was an attempt by China, a failed attempt, that Yangtse, there is a place where it is disputed and only held by India, only disputed territory held by India. They, unprecedentedly, they tried their last time there. An attempt, unprecedented attempt. While this happened for two months, through some of the systems, we could see that there is some issue which is building up. But finally, we could, get predict the time when they are coming with us. And the prediction helped us to place our resources at the right places. And we could foil the attempt of Chinese unprecedented attempt. And thereafter, having no casualties by us, and we could place our resources and evacuation in time. So, no casualty and highest, I think, unprecedented award we got in Section 19 of them. So, AI prediction tools can actually help you operate anywhere effectively. I would also say that we should also build agentic systems for higher fusion. And at national level also have all the systems of economic security and other things also deployed together. So that it can help our apex level leadership to decide that what are the actions they are taking, what are the second and third order effects of those actions, rather than looking at in isolation. The civil-military fusion is the order of the day. I think this gives responsible proactivity. And we can really enhance our capabilities there. Boost our talent. And lastly, more importantly, when we become this as a superpower, which we are aspiring, and I am sure we will be on that path. We, I am sure we are going to become a very responsible AI power. As we are a nuclear responsible nuclear power when we acquired the nuclear capability. No first use itself shows that India is a very responsible power. And this kind of super intelligence, once we acquire, would be very effectively and responsibly used. Thank you. Thank you, sir, for those valuable insights. We now move to our next keynote speaker, a distinguished civil servant, who has played a foundational role in shaping India's digital and AI journey. Shri Alkesh Kumar Sharma, Senior IAS Officer of the Kerala Cadre, served as Secretary, Ministry of Electronics and Information Technology, Government of India. Shri Alkesh Kumar Sharma is currently a member of the Public Enterprises Selection Board, Government of India, responsible for selecting top executives for Central Public Sector Enterprises. He earlier served as Secretary, Ministry of Electronics and Information Technology (MeitY), where he led major national digital initiatives including Digital Governance, India Stack, Global Digital Public Infrastructure, Bhashini, Digital India Language Translation Platform, the National Program on AI, Semiconductor Program, Production Linked Incentive schemes for electronics manufacturing, and the Digital Personal Data Protection Act, significantly shaping India's digital economy. Prior to this, he served as Secretary and Additional Secretary in the Cabinet Secretariat under the Prime Minister of India, where he appraised key cabinet proposals across infrastructure, finance, agriculture, industry, and urban development, and monitored mega national projects through the PMO's Project Monitoring Group. May I have Shri Alkesh Kumar Sharma ji on stage, please? Yes. Thank you, Lieutenant General Rana, Dr. Chandrika Kaushik, Mr. Subhash. Distinguished officials, my colleagues in MeitY, Dr. Sanjay Bahal, distinguished friends. It is a matter of great privilege for me to be here this morning. And at the outset, I would like to congratulate both Bell and Globals for organizing this event, which is quite appropriate, part of the Global AI Summit. And last year, I remember when Prime Minister, our Prime Minister was in the third AI Global Summit in Paris, he underscored the transformative impact of AI on society, emphasizing both its rapid evolution and widespread adoption. He highlighted AI expanding all in shaping politics, economics, national security, and governance. As we all know, artificial intelligence is no longer a technological advancement; it is a potential. It has the potential to drive innovation, enhance productivity, and act as a force multiplier for security in the defense. It is the strategic capability shaping economic power, geopolitical influence, and national resilience. In fact, AI is becoming the defining capability of the 21st century, reshaping competitiveness and sovereignty. And the decisions which we take today on AI investment and governance will go a long way in defining the trajectory of nations for decades. And that is why India has taken some early lead. And it was not long ago. Let me take a step back that we started talking about AI. It was in late 21 or early 22 when we said there is something called AI and we should start doing something about it. And that is when the first AI platform India was set up. We started the first National Program on AI in 2022 as part of Digital India initiative. And that is when when we launched the India Stack in 2022. This Digital India initiative actually helped India in strengthening its AI policy architecture and digital governance framework. There was a push towards trusted digital public infrastructure and indigenous technology ecosystem. It gained momentum. We developed institutional framework for structured AI adoption across sectors, AI in health, education, agriculture. The Center of Excellence system started announcing the budget and we started working on that. Develop national greater emphasis placed on data governance, DPI, innovative ecosystem and public-private partnership. And this has all, this has led to a strong base for India's AI mission which was launched a year later after we actually held the Global Partnership Summit in December 2023. A collaborative ecosystem between government, startup, academy and industry was set up. A framework for safe, trusted, secure and accountable digital transformation was set up. The foundations built during this period continue to support India's expanding AI ecosystem. Today, while developing the policy framework, we had no doubt in our mind that only a sovereign AI would help India in achieving a safe, trusted, secure and responsible AI. As we all know, rented models will not only have biases but data quality, security and Indian context will be missing in those. India's sovereign AI framework will develop, control and regulate AI systems aligned with our national interest. And it is not isolation; it is strategic autonomy combined with trusted collaboration. Today, India is witnessing AI-driven transformation across sectors, whether it is legal, finance, healthcare, agriculture, governance, manufacturing. You talk about it. While I am now in my new role picking up top executives for the CPSC in India, I am finding that CPs are developing new AI applications which are not only increasing efficiency in operations but they are creating higher value addition across the value chain from incremental to manifold jumps. As startups are disrupting traditional sectors, large enterprises are embedding AI operations. As we know, our policy framework, its, rests on seven important pillars. But let me come back to the AI in defense and strategic. India recognizes the strategic importance of AI in defense. It recognizes the need to leverage AI to enhance combat capabilities, streamline decision-making processes and optimize resource deployment. And that's essential for all our defense forces. In this context, AI offers immense force multiplier potential, enabling superior situational awareness, autonomous systems and data-driven strategies. As technology world must be aware of the latest what is happening in the to maintain our operational superiority in defense and national security context. India's AI policy framework must ensure control over critical data sets, reduce dependence on external AI systems, alignment with national doctrine and values, and protection of our strategic infrastructure. Now, AI is embedded across defense systems which Lieutenant General has talked about. But AI in defense must be controlled, auditable and accountable. It can create havoc if it is not properly controlled. Responsible AI in defense requires human oversight, ethical safeguards, risk assessment framework and clear accountability mechanisms. So, India must pursue sovereign AI with safety, trust and accountability, strategic capability with ethical oversight and indigenous innovation with global cooperation. India has the opportunity to become a leader in AI. We can lead responsibly, strategically and confidently. Thank you. Jai Hind. Thank you, sir. We now welcome one of India's most accomplished and respected defense scientists. She is an inspiration to many, especially women aspiring to lead in science, technology, and national security. Featured by India Today as a leading women defense leader, she has made outstanding contributions to India's strategic capabilities. It is none other than Dr. Chandrika Kaushik, Distinguished Scientist and Director General, Production Coordination and Services Interaction, DRDO. Let's welcome Chandrika Kaushik, ma'am, please. Thank you. Lieutenant General Rana, CNC, SFC, Mr. Alkesh Kumar Sharma, former Secretary, MeitY, Mr. Subhash Gopinath, ladies and gentlemen. First and foremost, I must compliment the Bell and Global team for organizing this very, very relevant topic, at least from the security agencies' perspective, on AI for strategic defense as a part of the AI Impact Summit 2026, which is absolutely imperative today, looking at the way things are changing.

The changing operational environments for soldiers are progressively becoming complex with the incorporation of advanced technologies into military operations, be it network-centric warfare, multi-domain operations, or unmanned combat systems. These complexities are resulting in increased information overloads, reduced reaction times, and quick decision-making for the human operator.

In order to cater to these challenges, it is the need of the hour to introduce autonomy into our military systems to augment the capabilities of the soldier. In June 2018, the Government of India published the National Strategy for AI, focusing on how India can leverage the transformative AI technologies to ensure inclusive growth in all sectors of society, from sensing to comprehension to action.

Today, AI has pervaded numerous end-user applications, ranging from computer vision to audio processing, natural language processing, knowledge representation, machine learning, expert systems, etc. The AI is being classified into three variants: descriptive AI, which describes what has happened; predictive, which is anticipating what is going to happen; and prescriptive AI, which recommends what to do.

With the emerging technological capabilities, AI is geared up to serve as a promising tool to provide enhanced situational awareness to the commanders and operators in highly contested battlefield environments, making our military systems more accurate and our military operations more swift and precise.

If we look at the various domains where AI will get integrated, some of these areas are: the ecosystem is working on underwater surveillance, navigation controls, advanced material characterization, hybrid high-fidelity models in the naval systems domain, trajectory and manual planning, target identification, threat classification, aerodynamics, and controls. So, I have listed out. I think for paucity of time, I will not be able to get into the details. But if you look at it, you look at armaments, combat engineering, missiles, strategic systems, electronics and communication systems, computational, cyber systems, everywhere there is an application where AI is getting seamlessly integrated.

Similarly, in the counter-unmanned aerial systems grid, it can be used to neutralize threats from sub-conventional aerial platforms like drones, loitering munitions, swarm drones, with the network system of AI-powered multi-layer detection and identification, dynamic resource allocation, tracking and engagement of targets through hard kill and soft kill measures. C4ISR operations with man-machine teaming in urban environments can be carried out using coordinated combat operations in complex, high-density environments where challenges of limited visibility, civilian presence, multi-story buildings, and ambush risks. All these can be handled by using AI-enabled advanced surveillance and integrated systems for enhancing the soldier's effectiveness, leading to very, very effective man-machine teaming.

In defense, reliability and robustness are critical system requirements to prevent mission failure and unintended consequences. Ensuring that AI applications are safe, resilient to attacks, and aligned with ethical values is imperative for our defense forces. In order to address this, DRDO has identified five core trustworthy AI principles: reliability and robustness, transparency, fairness, privacy, safety, and security, which address various vulnerabilities experienced by AI systems during different lifecycle stages of development. The trustworthiness principles have been further augmented with 11 qualifying criteria, and the implementation process has been duly published as a set of two policy documents, namely, "Evaluating Trustworthiness in AI Systems: The Framework Document" and "The Guidelines Document."

Indian military leadership will need interpretable AI decisions with explainability, human trust factors, and ethical considerations in local operating scenarios. The domestic framework is aligned with existing Indian policies and standards published by DGQA, SEMILIB, BIS. All these will be required to be alive. This framework has been officially released last month and is available for the relevant stakeholders and is geared towards providing a unified approach to bringing resilience and trustworthiness in AI-powered military systems.

Armed forces may need to institutionalize AI safety requirements as part of the acquisition lifecycle process, and the Directorate of Standardization is expected to define and enforce the minimum mandatory V&V protocols for operational scenarios. We expect that NIC will be required to build infrastructure for defense-grade secure data repositories for training, validation, and testing. The roadmap towards self-reliance in defense AI lies in establishing a national AI defense strategy, and General Rana spoke about it, detailing various aspects of autonomy, decision support, threat analysis, creating dedicated structures for integrating all the stakeholders in the ecosystem, using the ethical guidelines for AI, and the domestic data infrastructure will need to be ramped up to ensure secure and federated data spanning ISR sensors, battlefield, and historical combat data.

We will also need to work on labeling and annotation standards along with standardized processes which will need to be adopted across platforms and operational scenarios. A clear policy on sovereign data should be formulated and implemented at the earliest to prevent any dependency on foreign data and to ensure data residency, specifically for the defense domain. The model designs must cater to explainability, adversarial robustness, and fail-safe mechanisms, and dedicated indigenous protocols for domestic combat scenarios should be adopted for verification and validation.

I will just finish in another half a minute. Low-latency AI interface models, hardware accelerators, and trusted environments will be required for deployment and as enablers for resilience. Cyber hardening will be required for ensuring resilience to spoofing, jamming, and cyber attacks. It is proposed that adaptive command support may be adopted, wherein AI must augment and not replace human decision-making in mission-critical tasks, and a redundant human-in-the-loop conventional override system be included for backup and failover systems. Together, I am sure we can use AI very, very effectively and efficiently in the defense domain for building strategic capability. Thank you. Jai Hind.

Now, when we talk about non-proliferation of these capabilities in the nuclear domain, is it possible? Presently, not very successful, but we could have some non-proliferation worldwide so that non-state actors do not acquire them. But the same would be quite difficult because of obvious reasons when we talk about superintelligence, once it comes into being. Now, when we talk about this artificial intelligence use in the strategic domain and national security, there is a lot of usage of the same. In my present appointment as Director General DIA, I have been exposed to many of the open-source intelligence systems. If they combine well with the sensors of all the three services, cyber and space, and also some of the other elements like what NASA has developed, the artificial light, the light analysis software, you can really identify the hotspots emerging in your neighborhood or in your own area. So that kind of thing, which is already in place in certain places, can be further developed by us in times to come.

Now, when we talk about other usages, I think you are quite aware, right from the basic usage of logistics and preventing and maintaining your systems from the AI software. We are transitioning into command and control systems, improving our maritime domain awareness. At tactical levels, electronic warfare jamming, very effectively we can do that. And what we have seen in Operation Sindoor, the AI can now be very effective in combating misinformation, which is in vast numbers. We could see in Operation Sindoor.

Now, the question is, what we need to do? We need to do, and as India, when we talk about these capabilities. First of all, I would suggest that we should have, we should develop an AI security doctrine, and we should use our resources where actually, as per doctrine, where it is required more. And second is, if you take an example from the Russia-Ukraine conflict, Russia, there are a lot of systems which are getting deployed without waiting from top-down capabilities. Battlefield, low-cost AI-based systems, which are easily scalable and are cheap, are giving good results. They are increasing the kill chains of these autonomous systems and drones. So this needs to be deployed even while we work on the higher systems of artificial intelligence.

Now, taking some examples back, when I was commanding a corps in 4 Corps in Arunachal, where I was responsible for the security of the Line of Actual Control of Tawang and areas around, I remember we had a very, we had developed a small artificial intelligence software, low-ability, but when there was an attempt by China, a false attempt, that Yangtse, there is a place where it is disputed and only held by India, only disputed territory held by India, they, unprecedentedly, they tried their last time. There was an unprecedented attempt. While this happened for two months, through some of the systems, we could see that there is some issue which is building up. But finally, we could predict the time when they are coming with us. And the prediction helped us to place our resources at the right places. And we could foil the attempt of the Chinese unprecedented attempt. And thereafter, having no casualties by us, and we could place our resources and evacuation in time. So no casualties and the highest, I think, unprecedented guarantee award we got in Section 19 of them. So artificial intelligence prediction tools can actually help you operate anywhere effectively.

I would also say that we should also build agentic systems for higher fusion and at the national level also have all the systems of economic security and other things also deployed together. So that it can help our apex-level leadership to decide that what are the actions they are taking, what are the second and third-order effects of those actions, rather than looking at them in isolation. The civil-military fusion is the order of the day. I think this gives our responsible activity, and we can really enhance our capabilities there, boost our talent. And lastly, more importantly, when we become this, as a superpower which we are aspiring, and I am sure we will be on that path, I am sure we are going to become a very responsible AI power, as we are a nuclear responsible nuclear power. When we acquired the nuclear capability, our "no first use" itself shows that India is a very responsible power. And this kind of superintelligence, once we acquire, would be very effectively and responsibly used. Thank you.

Thank you, sir, for those valuable insights. We now move to our next keynote speaker, a distinguished civil servant who has played a foundational role in shaping India's digital and AI journey. Shri Alkesh Kumar Sharma, a senior IAS officer of the Kerala cadre, served as Secretary, Ministry of Electronics and Information Technology, Government of India. Shri Alkesh Kumar Sharma is currently a Member of the Public Enterprises Selection Board, Government of India, responsible for selecting top executives for Central Public Sector Enterprises. He earlier served as Secretary, Ministry of Electronics and Information Technology (MeitY), where he led major national digital initiatives, including Digital Governance, India Stack, Global Digital Public Infrastructure, Bhashini, Digital India Language Translation Platform, the National Program on AI, Semiconductor Program, Production Linked Incentive schemes for electronics manufacturing, and the Digital Personal Data Protection Act, significantly shaping India's digital economy. Prior to MeitY, he served as Secretary and Additional Secretary in the Cabinet Secretariat under the Prime Minister of India, where he appraised key cabinet proposals across infrastructure, finance, agriculture, industry, and urban development, and monitored mega national projects through the PMO's Project Monitoring Group. May I have Shri Alkesh Kumar Sharma ji on the stage, please?

Yes, thank you. Lieutenant General Rana, Dr. Chandrika Kaushik, Mr. Subhash, distinguished officials, my colleague in MeitY, Dr. Sanjay Bahal, distinguished friends. It is a matter of great privilege for me to be here this morning. And at the outset, I would like to congratulate both Bell and Globals for organizing this event, which is quite appropriate as part of the Global AI Summit. And last year, I remember when Prime Minister, our Prime Minister, was in the third AI Global Summit in Paris, he underscored the transformative impact of AI on society, emphasizing both its rapid evolution and widespread adoption. He highlighted AI expanding all in shaping politics, economics, national security, and governance.

Today, while developing the policy framework, we had no doubt in our mind that only a sovereign AI would help India in achieving a safe, trusted, secure, and responsible AI. As we all know, rented models will not only have biases, but data quality, security, and Indian context will be missing in those. India's sovereign AI framework will develop, control, and regulate AI systems aligned with our national interests. And it is not isolation; it is strategic autonomy combined with trusted collaboration.

Today, India is witnessing AI-driven transformation across sectors, whether it is legal, finance, healthcare, agriculture, governance, manufacturing. You talk about it. While I am now in my new role, picking up top executives for the CPSC in India, I am finding that CPs are developing new AI applications which are not only increasing efficiency in operations, but they are creating higher value addition across the value chain from incremental to manifold gains. As startups are disrupting traditional sectors, large enterprises are embedding AI operations. As we know, our policy framework, its, rests on seven important pillars.

But let me come back to the AI in defense and strategic endurance. India recognizes the strategic endurance of AI in defense. It recognizes the need to leverage AI to enhance combat capabilities, streamline decision-making processes, and optimize resource deployment. And that's important for all our defense forces. In this context, AI offers immense force multiplier potential, enabling superior situational awareness, autonomous systems, and data-driven strategies. As technology, world must be aware of the latest but is happening in the to maintain our operational superiority in defense and national security context. India's AI policy framework must ensure control over critical data sets, reduce dependence on external AI systems, alignment with national doctrine and values, and protection of our strategic infrastructure.

Now, AI is embedded across defense systems, which Lieutenant General has talked about. But AI in defense must be controlled, auditable, and accountable. It can create havoc if it is not properly controlled. Responsible AI in defense requires human oversight, ethical safeguards, risk assessment frameworks, and clear accountability mechanisms. So, India must pursue sovereign AI with safety, trust, and accountability, strategic capability with ethical oversight, and indigenous innovation with global cooperation. India has the opportunity to become a leader in AI. We can lead responsibly, strategically, and confidently. Thank you. Jai Hind.

Thank you, sir. We now welcome one of India's most accomplished and respected defense scientists. She is an inspiration to many, especially women aspiring to lead in science, technology, and national security. Featured by India Today as a leading women defense leader, she has made outstanding contributions to India's strategic capabilities. It is none other than Dr. Chandrika Kaushik, Distinguished Scientist and Director General, Production Coordination and Services Interaction, DRDO. Let's welcome Chandrika Kaushik, ma'am, please.

Thank you. Lieutenant General Rana, CNC, SFC, Mr. Alkesh Kumar Sharma, former Secretary MeitY, Mr. Subhash Gopinath, ladies and gentlemen. First and foremost, I must compliment the Bell and Global team for organizing this very, very relevant topic, at least from the security agency perspective, on AI for Strategic Defense, as part of the AI Impact Summit 2026, which is absolutely imperative today, looking at the way things are changing. Operational environments for soldiers are progressively becoming complex with the incorporation of advanced technologies into military operations, be it network-centric warfare, multi-domain operations, unmanned combat systems. These complexities.

The complex operational environments for soldiers are progressively becoming more complex with the incorporation of advanced technologies into military operations, be it network-centric warfare, multi-domain operations, or unmanned combat systems. These complexities are resulting in increased information overloads, reduced reaction times, and quick decision cycles for the human operator. In order to cater to these challenges, it is the need of the hour to introduce autonomy into our military systems to augment the capabilities of the soldier.

In June 2018, the Government of India published the National Strategy for AI, focusing on how India can leverage the transformative AI technologies to ensure inclusive growth in all sectors of society. From sensing to comprehension to action. Today, AI has pervaded numerous and user applications ranging from computer vision to audio processing, natural language processing, knowledge representation, machine learning, expert systems, etcetera. AI is being classified into three variants: descriptive AI, which describes what has happened; predictive, which is anticipating what is going to happen; and prescriptive AI, which recommends what to do. With the emerging technological capabilities, AI is geared up to serve as a promising tool to provide enhanced situational awareness to the commanders and operators in highly contested battlefield environments, making our military systems more accurate and our military operations more swift and precise.

If we look at the various domains where AI will get integrated, some of these areas are: the ecosystem is working on underwater surveillance, navigation controls, advanced material characterization, hybrid high-fidelity models in the naval system domain, trajectory and maneuver planning, target identification, threat classification, aerodynamics, and controls. So, I have listed out. I think for paucity of time, I will not be able to get into the details. But if you look at it, you look at armaments, combat engineering, missiles, strategic systems, electronics and communication systems, computational cyber systems, everywhere there is an application where AI is getting seamlessly integrated.

In the near future, AI will find applications in multiple operational areas. And some of the use cases I am just going to deliberate in the next few minutes: with the upcoming challenges of long-range interceptions, reduced reaction times, and saturation of airspace, AI-based decision support systems can greatly augment faster, adaptive, and intelligent decision-making across early surveillance, automated threat classification, quick and autonomous interception, and real-time battle damage assessment. Similarly, in the counter-unmanned aerial systems grid, it can be used to neutralize threats from sub-conventional aerial platforms like drones, loitering munitions, swarm drones, with the network system of AI-powered multi-layer detection and identification, dynamic resource allocation, tracking and engagement of targets through hard kill and soft kill measures. C4ISR operations with man-machine teaming in urban environments can be carried out using coordinated combat operations in complex, high-density environments where challenges of limited visibility, civilian presence, multi-storied buildings, and ambush risks exist. All these can be handled by using AI-enabled advanced surveillance and integrated systems for enhancing the soldier's effectiveness, leading to very, very effective man-machine teaming.

With increasing physical, psychological, and cognitive loads in highly contested environments, the future soldier and his equipment will need to be integrated with all nearby sensors and equipment. In such a scenario, AI will help in integrating the sensory information along with the commander's decision-making process through shared mental models. AI will also help in modernizing and coordinating India's integrated EW capabilities across services with indigenous networks and real-time AI-powered electronic warfare suits like smart jammers, cognitive radios, and cyber-secure infrastructure. Weight and size optimization for performance enhancement and operations in extreme environments call for R&D on new-age materials where AI-based modeling and simulations will help in synthesizing these materials, providing higher accuracy, help with adaptive manufacturing, predictive quality control for high-fidelity models, and so on and so forth.

In defense, reliability and robustness are critical system requirements to prevent mission failure and unintended consequences. Ensuring that AI applications are safe, resilient to attacks, and aligned with ethical values is imperative for our defense forces. In order to address this, DRDO has identified five core trustworthy AI principles: reliability and robustness, transparency, fairness, privacy, safety, and security, which address various vulnerabilities experienced by AI systems during different life cycle stages of development. The trustworthiness principles have been further augmented with 11 qualifying criteria, and the implementation process has been duly published as a set of two policy documents, namely "Evaluating Trustworthiness in AI Systems: The Framework Document" and "The Guidelines Document."

As we know that AI is as good as it is trained, it heavily depends on training data, most of which is taken from the open source. With respect to India, with its varied geography, it essentially needs domestic military data infrastructure for training, validation, and testing data sets. Imported V&V tools may have vulnerabilities. Also, the defense models will be classified and sensitive and need to be validated in indigenous environments. An indigenous AI verification and validation framework will ensure that all data remains within our sovereign boundaries and Indian control, adhering to domestic policies and regulations. Indian military leadership will need interpretable AI decisions with explainability, human trust factors, and ethical considerations in local operating scenarios. The domestic framework is aligned with existing Indian policies and standards published by DGQA, CEL, BIS. All these will be required to be aligned. This framework has been officially released last month and is available for the relevant stakeholders and is geared towards providing a unified approach to bringing resilience and trustworthiness in AI-powered military systems.

The roadmap towards self-reliance in defense AI lies in establishing a national AI defense strategy. And General Rana spoke about it, detailing various aspects of autonomy, decision support, threat analysis, creating dedicated structures for integrating all the stakeholders in the ecosystem, using the ethical guidelines for. And the domestic data infrastructure will need to be ramped up to ensure secure and federated data spanning ISR sensors, battlefield, and historical combat data. We will also need to work on labeling and annotation standards along with standardized processes, which will need to be adopted across platforms and operational scenarios. A clear policy on sovereign data should be formulated and implemented at the earliest to prevent any dependency on foreign data and to ensure data residency, specifically for the defense domain. The model design must cater to explainability, adversarial robustness, and fail-safe mechanisms, and dedicated indigenous protocols for domestic combat scenarios should be adopted for verification and validation.

I will just finish in another half a minute. Low latency AI interface models, hardware accelerators, and trusted environments will be required for deployment and edge enablement for resilience. Cyber hardening will be required for ensuring resilience to spoofing, jamming, and cyber attacks. It is proposed that adaptive command support may be adopted, wherein AI must augment and not replace human decision-making in mission-critical tasks, and a redundant human-in-the-loop conventional override system be included for backup and failover systems. Together, I am sure we can use AI very, very effectively and efficiently in the defense domain for building strategic capability. Thank you. Jai Hind.

Thank you, ma'am. That was truly inspiring. Today's keynotes from all three dignitaries have been extremely insightful. My sincere thanks to Lieutenant General Dinesh Singh Rana, Shri Alkesh Kumar Sharma, and Dr. Chandrika Kaushik for sharing valuable perspectives. That may play also a role in strategic deterrence. As we are talking about nuclear deterrence, we have seen in the Cold War, the mutual assured destruction. Now, there is a talk about the mutual assured AI malfunction, which has already been talked about, and this may be actually a reality. ASI, that is superintelligence, having a kind of tremendous power, could also be used by a country which possesses this capability against a weaker country, in the form of what we have seen in the nuclear domain: nuclear brinkmanship or nuclear blackmail. Now, when we talk about non-proliferation of these capabilities in the nuclear domain, it is possible. Presently, not very successful, but we could have some non-proliferation worldwide, so that non-state actors don't acquire them. But the same would be quite difficult because of obvious reasons when we talk about superintelligence once it comes into being.

Now, when we talk about this artificial intelligence in the strategic domain and national security, there is a lot of usage of the same. In my present appointment as Director General DIA, I have been exposed to many open-source intelligence systems. If they combine well with the sensors of all the three services, cyber and space, and also some of the other elements like what NASA has developed, the artificial light, the light analysis software, you can really identify the hotspots emerging in your neighborhood or in your own area. So that kind of thing, which is already in place in certain places, can be further developed by us in times to come. Now, in cyber, when we talk about AI, it inherently is more offensive in nature. In cyber, we have seen the GTG 102 attack which happened in September last year, the autonomous cyber attacks by Chinese state actors against the US systems. But at the same time, the cyber capabilities can be very effectively used to identify the vulnerabilities and plug the gaps.

Now, when we talk about other usages, I think you are quite aware, right from the basic usage of logistics and preventing and maintaining your systems from the AI software, we can. We are transitioning into command and control systems, improving our maritime domain awareness at tactical levels, electronic warfare jamming, very effectively we can do that. And what we have seen in Operation Sindoor, the AI can now be very effective in combating misinformation, which is in vast numbers. We saw in Operation Sindoor. Now, the question is what we need to do. We need to do, and as India, when we talk about the capabilities, first of all, I would suggest that we should have, we should develop an AI security doctrine, and we should use our resources where actually, as per the doctrine, where it is required more. And second is, if you take an example from the Russia-Ukraine conflict, Russia, there are a lot of systems which are getting deployed without waiting from top-down capabilities. Battlefield, low-cost AI-based systems which are easily scalable and are cheap are getting.

Shri Subhash Gopinath, ladies and gentlemen, first and foremost, I must compliment the BEL and Global team for organizing this very, very relevant topic, at least from the security agencies' perspective, on AI for strategic defense as part of the AI Impact Summit 2026, which is absolutely imperative today, looking at the way things are changing.

Of all the three services, cyber and space, and also some of the other elements like what NASA has developed, the artificial light, the light analysis software, you can really identify the hotspots emerging in your neighborhood or in your own area. So that kind of thing, which is already in place in certain places, can be further developed by us in times to come.

Now, in cyber, when we talk about AI, it inherently is more offensive in nature. In cyber, we have seen the GTG 102 attack which happened in September last year, the autonomous cyber attacks by Chinese state actors against the US systems. But at the same time, the cyber capabilities can be very effectively used to identify the vulnerabilities and plug the gaps.

Now, when we talk about other usages, I think you are quite aware, right from the basic usage of logistics and prevent and maintaining your systems from the AI software, we can, we are improving command and control systems, improving our maritime domain awareness at tactical levels, electronic warfare jamming, very effectively we can do that. And what we have seen in operations, the AI can now be very, very effective in combating misinformation, which is in vast numbers. We can, we saw in operations Sindoor.

Now, the question is what we need to do. We need to do, and as India, when we talk about the capabilities. First of all, I would suggest that we should have, we should develop an AI security doctrine, and we should use our resources where actually, as per doctrine, where it is required more.

And second is, if you take an example from the Russia-Ukraine conflict, the Russia, there are a lot of systems which are getting deployed without waiting from top-down capabilities, battlefield low-cost AI-based systems which are easily scalable and are cheap, are giving good results. They are increasing the kill chains of these autonomous systems and drones. So, this needs to be deployed, while we work on the higher systems of artificial intelligence.

Now, taking some examples back, when I was commanding a corps in IV Corps in Arunachal, where I was responsible for security of Line of Actual Control of Tawang and area around. I remember we had a very, we had developed a small artificial intelligence software, low-ability. But when there was an attempt by China, a failed attempt, that Yangtse, there is a place where it is disputed and only held by India, only disputed territory held by India, they, unprecedentedly, they tried their last time, their attempt, unprecedented attempt. While this happened for two months, through some of the systems, we could see that there is some issue which is building up. But finally, we could predict the time when they are coming with us. And the prediction helped us to place our resources at the right places. And we could foil the attempt of Chinese unprecedented attempt. And thereafter, having no casualties by us, and we could place our resources and evacuation in time. So, no casualties and highest, I think, unprecedented gallantry award we got in Section 19 of them. So, AI prediction tools can actually help you operate any very effectively.

I would also say that we should also build agentic systems for higher fusion and at national level also have all the systems of economic security and other things also deployed together so that it can help our apex level leadership to decide that what are the actions they are taking, what are the second and third order effects of those actions, rather than looking at in isolation. The civil-military fusion is the order of the day. I think this gives responsible, proactive activity and we can really enhance our capabilities there. Boost our talent and lastly, more importantly, when we become this as a superpower which we are aspiring and I am sure we will be on that path. I am sure we are going to become a very responsible AI power. As we are a responsible nuclear power when we acquired the nuclear capability, no first use itself shows that India is a very responsible power. And this kind of superintelligence, once we acquire, would be very effectively and responsibly used. Thank you.

Thank you, sir, for those valuable insights. We now move to our next key speaker, a distinguished civil servant who has played a foundational role in shaping India's digital and AI journey, Shri Alkesh Kumar Sharma, Senior IAS Officer of the Kerala cadre, served as Secretary, Ministry of Electronics and Information Technology, Government of India. Shri Alkesh Kumar Sharma is currently a Member of the Public Enterprises Selection Board, Government of India, responsible for selecting top executives for Central Public Sector Enterprises. He earlier served as Secretary, Ministry of Electronics and Information Technology (MeitY), where he led major national digital initiatives including Digital Governance, India Stack, Global Digital Public Infrastructure, Bhashini, Digital India Language Translation Platform, the National Program on AI, Semiconductor Program, Production Linked Incentive schemes for electronics manufacturing, and the Digital Personal Data Protection Act, significantly shaping India's digital economy priorities. He served as Secretary and Additional Secretary in the Cabinet Secretariat under the Prime Minister of India, where he approved key cabinet proposals across infrastructure, finance, agriculture, industry, and urban development, and monitored mega national projects through the PMO's Project Monitoring Group. May I have Shri Alkesh Kumar Sharma Ji on stage, please?

Yes, thank you. Lieutenant General Rana, Dr. Chandraka Kaushik, Mr. Subhash, distinguished officials, my colleagues in MeitY, Dr. Sanjay Bahl, distinguished friends. It is a matter of great privilege for me to be here this morning. And at the outset, I would like to congratulate both BEL and Globals for organizing this event, which is quite appropriate as part of the Global AI Summit. And last year, I remember when Prime Minister, our Prime Minister, was in the third AI Global Summit in Paris, he underscored the transformative impact of AI on society, emphasizing both its rapid evolution and widespread adoption. He highlighted AI expanding its role in shaping politics, economics, national security, and governance. As we all know, artificial intelligence is no longer a technological advancement; it is a potential. It has the potential to drive innovation, enhance productivity, and act as a force multiplier for security in the defense. It is the strategic capability shaping economic power, geopolitical influence, and national resilience. In fact, AI is becoming the defining capability of the 21st century, reshaping competitiveness and sovereignty. And the decisions which we take today on AI investment and governance will go a long way in defining the trajectory of nations for decades. And that is why India has taken some early lead. And it was not long ago. Let me take a step back that we started talking about AI. It was in late 21 or early 22 when we said there is something called AI and we should start doing something about it. And that is when the first AI platform, India, was set up. We started the first National Program on AI in 2022 as part of the Digital India initiative. And that is when, when we launched the India Stack in 2022, this Digital India initiative actually helped India in strengthening its AI policy architecture and digital governance framework. There was a push towards trusted digital public infrastructure and indigenous technology ecosystem. It gained momentum. We developed institutional frameworks for structured AI adoption across sectors, AI in health, education, agriculture. The Centers of Excellence started, announced in the budget, and we started working on that. Developed national grids, placed emphasis on data governance, DPI, innovative ecosystem, and public-private partnership. And this has all, this has led to a strong base for India's AI mission, which was launched a year later after we actually held the Global Partnership Summit in December 2023. A collaborative ecosystem between government, startups, academia, and industry was set up. A framework for safe, trusted, secure, and accountable digital transformation was set up. The foundations built during this period continue to support India's expanding AI ecosystem. Today, while developing the policy framework, we had no doubt in our mind that only a sovereign AI would help India in achieving a safe, trusted, secure, and responsible AI. As we all know, rented models will not only have biases, but data quality, security, and Indian context will be missing in those. India's sovereign AI framework will develop, control, and regulate AI systems aligned with our national interests. And it is not isolation; it is strategic autonomy combined with trusted collaboration. Today, India is witnessing AI-driven transformation across sectors, whether it is legal, finance, healthcare, agriculture, governance, manufacturing. I am sure you talk about it. And while I am now in my new role, picking up top executives for the CPSEs in India, I find that CPSEs are developing new AI applications which are not only increasing efficiency in operations, but they are creating higher value addition across the value chain, from incremental to manifold jumps. As startups are disrupting traditional sectors, large enterprises are embedding AI operations. As we know, our policy framework, its, rests on seven important pillars. But let me come back to the AI in defense and strategic imperatives. India recognizes the strategic imperatives of AI in defense. It recognizes the need to leverage AI to enhance combat capabilities, streamline decision-making processes, and optimize resource deployment. And that's important for all our defense forces. In this context, AI offers immense force multiplier potential, enabling superior situational awareness, autonomous systems, and data defense strategies. As technology, the world must be aware of the latest what is happening in the two to maintain our operational superiority. In defense and national security context, India's AI policy framework must ensure control over our critical datasets, reduce dependence on external AI systems, alignment with national doctrine and values, and protection of our strategic infrastructure. Now, AI is embedded across defense systems, which Lieutenant General has talked about. But AI in defense must be controlled, auditable, and accountable. It can create havoc if it is not properly controlled. Responsible AI in defense requires human oversight, ethical safeguards, risk assessment frameworks, and clear accountability mechanisms. So, India must pursue sovereign AI with safety, trust, and accountability, strategic capability with ethical oversight, and indigenous innovation with global cooperation. India has the opportunity to become a leader in AI. We can lead responsibly, strategically, and confidently. Thank you. Jai Hind.

Thank you, sir. We now welcome one of India's most accomplished and respected defense scientists. She is an inspiration to many, especially women aspiring to lead in science, technology, and national security. Featured by India Today as a leading woman defense leader, she has made outstanding contributions to India's strategic capabilities. It is none other than Dr. Chandrika Kaushik, Distinguished Scientist and Director General, Production Coordination and Services Interaction, DRDO. Let's welcome Chandrika Kaushik Ma'am, please.

Thank you. Thank you, Lieutenant General Rana, CNCSFC, Mr. Alkesh Kumar Sharma, former Secretary MeitY, Mr. Subhash Gopinath, ladies and gentlemen. First and foremost, I must compliment the BEL and Global team for organizing this very, very relevant topic, at least from the security agency perspective, on AI for Strategic Defense, as part of the AI Impact Summit 2026, which is absolutely imperative today. Looking at the way things are changing, operational environments for soldiers are progressively becoming complex with the incorporation of advanced technologies into military operations, be it network-centric warfare, multi-domain operations, unmanned combat systems. The complexities are resulting in increased information overloads, reduced reaction times, and quick decision cycles for the human operator. In order to cater to these challenges, it is the need of the hour to introduce autonomy into our military systems to augment the capabilities of the soldier.

In June 2018, the Government of India published the National Strategy for AI, focusing on how India can leverage the transformative AI technologies to ensure inclusive growth in all sectors of the society, from sensing to comprehension to action. Today, AI has pervaded numerous end-user applications ranging from computer vision to audio processing, natural language processing, knowledge representation, machine learning, expert systems, etcetera. The AI is being classified into three variants: descriptive AI, which describes what has happened; predictive, which is anticipating what is going to happen; and prescriptive AI, which recommends what to do. With the emerging technological capabilities, AI is geared up to serve as a promising tool to provide enhanced situational awareness to the commanders and operators in highly contested battlefield environments, making our military systems more accurate and our military operations more swift and precise.

If we look at the various domains where AI will get integrated, and some of the areas are, the ecosystem is working on underwater surveillance, navigation controls, advanced material characterization, hybrid high-fidelity models in the naval systems domain, trajectory and maneuver planning, target identification, threat classification, aerodynamics and controls. So, I have listed out. I think for paucity of time, I will not be able to get into the details. But if you look at it, you look at arms, combat engineering, missiles, strategic systems, electronics and communication systems, computational cyber systems, everywhere there is an application where AI is getting seamlessly integrated.

In the near future, AI will find applications in multiple operational areas, and some of the use cases I am just going to deliberate in the next few minutes. With the upcoming challenges of long-range interceptions, reduced reaction times, and saturation of airspace, AI-based decision support systems can greatly augment faster, adaptive, and intelligent decision-making across early surveillance, automated threat classification, quick and autonomous interception, and real-time battle damage assessment. Similarly, in the counter-unmanned aerial systems grid, it can be used to neutralize threats from sub-conventional aerial platforms like drones, loitering munitions, swarm drones, with the network system of AI-powered multi-layer detection and identification, dynamic resource allocation, tracking and engagement of targets through hard kill and soft kill measures. C4ISR operations with man-machine teaming in urban environments can be carried out using coordinated combat operations in complex, high-density environments where challenges of limited visibility, civilian presence, multi-storied buildings, and ambush risks, all these can be handled by using AI-enabled advanced surveillance and integrated systems for enhancing the soldier's effectiveness, leading to very, very effective man-machine teaming. With increasing physical, psychological, and cognitive loads in highly contested environments, the future soldier and his equipment will need to be integrated with all nearby sensors and equipment. In such a scenario, AI will help in integrating the sensory information along with the commander's decision-making process through shared mental models. AI will also help in modernizing and coordinating the integrated India's EW capabilities across services with indigenous networks and real-time AI-powered electronic warfare suites like smart jammers, cognitive radios, and cyber-secure infrastructure. Weight and size optimization for performance enhancement and operations in extreme environments. Call for R&D on new-age materials where AI-based modeling and simulations will help in synthesizing the materials, providing higher accuracy, help with adaptive manufacturing, predictive quality control for high-fidelity models, and so on and so forth.

In defense, reliability and robustness is a critical system requirement to prevent mission failure and unintended consequences. Ensuring that AI applications are safe, resilient to attacks, and aligned with ethical values is imperative for our defense forces. In order to address this, DRDO has identified five core trustworthy AI principles, namely reliability and robustness, transparency, fairness, privacy, safety, and security, which address various vulnerabilities experienced by AI systems during different life cycle stages of development. The trustworthiness principles have been further augmented with 11 qualifying criteria, and the implementation process has been duly published as a set of two policy documents, namely, "Evaluating Trustworthiness in AI Systems: The Framework Document" and "The Guidelines Document." As we know that AI is as good as it is trained. It heavily depends on training data, most of which is taken from the open source. With respect to India, with its varied geography, essentially needs domestic military data infrastructure for training, validation, and testing data sets. Imported V&V tools may have vulnerabilities. Also, the defense models will be classified and sensitive and need to be validated in indigenous environments. And an indigenous AI verification and validation framework will ensure that all data remains within our sovereign boundaries and Indian control, adhering to domestic policies and regulations. Indian military leadership will need interpretable AI decisions with explainability, human trust factors, and ethical considerations in local operating scenarios. The domestic framework is aligned with existing Indian policies and standards published by DGQA, SEMILAC, BIS. All these will be required to be aligned. This framework has been officially released last month and is available for the relevant stakeholders and is geared towards providing a unified approach to bringing resilience and trustworthiness in AI-powered military systems.

Moving forward, we need a whole-of-nation approach for faster progression and utilization of AI-enabled solutions in the defense domain. It is our expectation that academia will establish validation and verification maturity models for each AI and explainability. And we expect industries to play a leading role in developing indigenous AI V&V toolkits, tool chains, including model probes, adversarial test beds, etcetera. DRDO can support in building standardized defense AI test suites for each domain and use case scenarios. And the quality assurance agencies need to institutionalize, need to formalize the AI V&V test labs and sandboxes for qualification and airworthiness. Armed forces may need to institutionalize AI safety requirements as part of the acquisition life cycle process. And Directorate of Standardization is expected to define and enforce the minimum mandatory V&V protocols for operational scenarios. We expect that NIC will be required to build infrastructure for defense-grade secure data repositories for training, validation, and tested. The roadmap towards self-reliance in defense AI lies in establishing a national AI defense strategy. And General Rana spoke about it, detailing various aspects of autonomy, decision support, threat analysis, creating dedicated structures for integrating all the stakeholders in the ecosystem, using the ethical guidelines for standards, and the domestic data infrastructure will need to be ramped up to ensure secure and federated data spanning ISR sensors, battlefield, and historical combat data. We will also need to work on labeling and annotation standards along with standardized processes which will need to be adopted across platforms and operational scenarios. A clear policy on sovereign data should be formulated and implemented at the earliest to prevent any dependency on foreign data and to ensure data residency, specifically for the defense domain. The model design must cater to explainability, adversarial robustness, and fail-safe mechanisms, and dedicated indigenous protocols for domestic combat scenarios should be adopted for verification and validation.

I will just finish in another half a minute. Low latency AI interface models, hardware accelerators, and trusted environments will be required for deployment and enablement for resilience. Cyber hardening will be required for ensuring resilience to spoofing, jamming, and cyber attacks. It is proposed that adaptive combat command support may be adopted, wherein AI must augment and not replace human decision-making in mission-critical tasks, and a redundant human-in-the-loop conventional override system be included for backup and failover systems. Together, I am sure we can use AI very, very effectively and efficiently in the defense domain for building strategic capability. Thank you. Jai Hind.

Thank you, Ma'am. That was truly inspiring. Today's keynotes from all three dignitaries have been extremely insightful. My sincere thanks to Lieutenant General Dinesh Singh Rana, Shri Alkesh Kumar Sharma, and Dr. Chandrika Kaushik for sharing valuable perspectives. That may play also a role in strategic deterrence. As we are talking about nuclear deterrence, we have seen in the Cold War, the Mutual Assured Destruction. Now, there is a talk about the mutual assured AI malfunction, which is already been talked about, and this may be actually a reality. AI, that is superintelligence, having a kind of tremendous power, could also be used by a country which possesses this capability against a weaker country, in the form of what we have seen in the nuclear domain, nuclear brinkmanship or nuclear blackmail. Now, when we talk about non-proliferation of these capabilities in the nuclear domain is possible, presently not very successful, but we could have some non-proliferation worldwide, so that non-state actors don't acquire them. But the same would be quite difficult because of obvious reasons, when we talk about superintelligence, once it comes into being. Now, when we talk about this artificial intelligence in the strategic domain and national security, there is a lot of usage of the same. In my present appointment as Director General DIA, I have been exposed to many of the open-source intelligence systems. If they combine well with the sensors of all the three services, cyber...

And space, and also some of the, uh, other elements like, like what NASA has developed, the artificial light, the light analysis software. You can really identify the hotspots emerging in your neighborhood or in your own area. So that kind of thing, which is already in place in certain places, can be further developed by us in times to come.

Now, uh, in cyber, when we talk about AI, inherently it is more offensive in nature. In cyber, we have seen the GTG 102 attack which happened in September last year, uh, the autonomous cyber attacks by Chinese state actors against the US, uh, systems. Uh, but at the same time, the cyber capabilities can be very effectively used to identify the vulnerabilities and plug the gaps.

Now, when we talk about other usages, I think you are quite aware, right from the basic usage of logistics and prevent and maintaining your systems from the AI software, we can, we are, uh, command and control systems, improving our maritime domain awareness at tactical levels, electronic warfare jamming, very effectively we can do that. And what we have seen in operations, the AI can now be very, very effective in combating misinformation, which is in vast numbers. We can, we saw in operations Sindoor.

Now, the question is, what we need to do. We need to do, and as, as India, uh, when we talk about the capabilities. First of all, I would suggest that we should have, we should develop an AI security doctrine, and we should use our resources where actually, as per doctrine, where it is required more.

And second is, if you take an example from the Russia-Ukraine conflict. Uh, Russia, there are a lot of systems which are getting deployed without waiting from top-down capabilities. Battlefield, low-cost, cheap AI-based systems which are easily scalable and are cheap, are getting giving good results. They are increasing the kill chains of these autonomous systems and drones. So, uh, these need to be deployed, uh, even as we work on the higher systems of artificial intelligence.

Now, taking some example back, when I was commanding a corps in Four Corps, uh, in Arunachal, where I was responsible for security of Line of Actual Control, uh, of Tawang and area around. I remember we had a very, we had developed a small artificial intelligence software, uh, low-ability. But, uh, when there was an attempt by China, a failed attempt, that Yangtse, there is a place where it is disputed and only held by India, only disputed territory held by India. Uh, they, uh, unprecedentedly, they tried their last time, their attempt. Unprecedented attempt. Uh, while this happened for two months, through some of the systems, we could see that there is some issue which is building up. But finally, we could, get predict the time when they are coming with us. And the prediction helped us to place our resources at the right places, and we could foil the attempt of Chinese unprecedented attempt. And thereafter, having no casualties, uh, by us, and we could place our resources and evacuation in time. So, no casualties, and highest, I think, unprecedented gallantry award we got in the section 19 of them. Uh, so, AI prediction tools can actually help you operate anywhere very effectively.

I would also say that we should also build agentic systems for higher fusion and at national level also have all the systems of economic security and other things also deployed together so that it can help our apex level leadership to decide that what are the actions they are taking, what are the second and third order effects of those actions, rather than looking at in isolation. Uh, the civil-military fusion is the order of the day. I think, uh, this gives, uh, responsible, uh, uh, activity, and we can really enhance our capabilities there. Boost our talent. And lastly, more importantly, when we become this, uh, as a superpower, which we are aspiring, and I am sure we will be on that path. Uh, we, I am sure we are going to become a very responsible AI power. As we are a nuclear responsible nuclear power, when we acquired the nuclear capability. Uh, no first use, itself shows that India is a very responsible power. And this kind of super intelligence, once we acquire, would be very effectively and responsibly used. Thank you.

Thank you, sir, for those valuable insights. We now move to our next key speaker, a distinguished civil servant who has played a foundational role in shaping India's digital and AI journey, Shri Alkesh Kumar Sharma, Senior IAS Officer of the Kerala Cadre, served as Secretary, Ministry of Electronics and Information Technology, Government of India. Shri Alkesh Kumar Sharma is currently a Member of the Public Enterprise Selection Board, Government of India, responsible for selecting top executives for Central Public Sector Enterprises. He earlier served as Secretary, Ministry of Electronics and Information Technology (MeitY), where he led major national digital initiatives including Digital Governance, India Stack, Global Digital Public Infrastructure, Bhashini, Digital India Language Translation Platform, the National Program on AI, Semiconductor Program, Production Linked Incentive schemes for electronics manufacturing, and the Digital Personal Data Protection Act, significantly shaping India's digital economy priorities. He served as Secretary and Additional Secretary in the Cabinet Secretariat under the Prime Minister of India, where he appraised key cabinet proposals across infrastructure, finance, agriculture, industry, and urban development, and monitored mega national projects through the PMO's Project Monitoring Group. May I have Shri Alkesh Kumar Sharma Ji on stage, please?

Thank you. Lieutenant General Rana, Dr. Chandrika Kaushik, Mr. Subhash, distinguished officials, my colleague in MeitY, Dr. Sanjay Bahal, distinguished friends. It is a matter of great privilege for me to be here this morning. And at the outset, I would like to congratulate both BEL and Globals for organizing this event, which is quite appropriate as part of the Global AI Summit. And last year, I remember when Prime Minister, our Prime Minister, was in the third AI Global Summit in Paris, he underscored the transformative impact of AI on society, emphasizing both its rapid evolution and widespread adoption. He highlighted AI expanding its role in shaping politics, economics, national security, and governance. As we all know, artificial intelligence is no longer a technological advancement; it is a potential. It has the potential to drive innovation, enhance productivity, and act as a force multiplier for security in the defense. It is the strategic capability shaping economic power, geopolitical influence, and national resilience. In fact, AI is becoming the defining capability of the 21st century, reshaping competitiveness and sovereignty. And the decisions which we take today on AI investment and governance will go a long way in defining the trajectory of nations for decades. And that is why India has taken some early lead, and it was not long ago. Let me take a step back that we started talking about AI. It was in late '21 or early '22 when we said there is something called AI and we should start doing something about it. And that is when the first AI platform, India, was set up. We started the first National Program on AI in 2022 as part of Digital India initiative. And that is when, when we launched the India Stack in 2022. This Digital India initiative actually helped India in strengthening its AI policy architecture and digital governance framework. There was a push towards trusted digital public infrastructure and indigenous technology ecosystem. It gained momentum. We developed institutional frameworks for structured AI adoption across sectors, AI in health, education, agriculture. The Centers of Excellence started announced in the budget, and we started working on that. Develop national greater emphasis placed on data governance, DPI, innovative ecosystem, and public-private partnership. And this has all, this has led to a strong base for India's AI mission, which was launched a year later after we actually held the Global Partnership Summit in December 2023. A collaborative ecosystem between government, startups, academia, and industry was set up. A framework for safe, trusted, secure, and accountable digital transformation was set up. The foundations built during this period continue to support India's expanding AI ecosystem. Today, while developing the policy framework, we had no doubt in our mind that only sovereign AI would help India in achieving a safe, trusted, secure, and responsible AI. As we all know, rented models will not only have biases, but data quality, security, and Indian context will be missing in those. India's sovereign AI framework will develop, control, and regulate AI systems aligned with our national interests. And it is not isolation; it is strategic autonomy combined with trusted collaboration. Today, India is witnessing AI-driven transformation across sectors, whether it is legal, finance, healthcare, agriculture, governance, manufacturing. I'm sure you talk about it. And while I am now in my new role, picking up top executives for the CPSUs in India, I am finding that CPSUs are developing new AI applications which are not only increasing efficiency in operations, but they are creating higher value addition across the value chain from incremental to manifold gains. As startups are disrupting traditional sectors, large enterprises are embedding AI operations. As we know, our policy framework, its, uh, rests on seven important pillars. But let me come back to the AI in defense and strategic intelligence. India recognizes the strength of AI in defense. It recognizes the need to leverage AI to enhance combat capabilities, streamline decision-making processes, and optimize resource deployment. And that's essential for all our defense forces. In this context, AI offers immense force multiplier potential, enabling superior situational awareness, autonomous systems, and data-driven strategies. As technology, the world must be aware of the latest what is happening in the two, to maintain our operational superiority. In defense and national security context, India's AI policy framework must ensure control over critical datasets, reduce dependence on external AI systems, alignment with national doctrine and values, and protection of our strategic infrastructure. Now, AI is embedded across defense systems, which Lieutenant General has talked about. But AI in defense must be controlled, auditable, and accountable. It can create havoc if it is not properly controlled. Responsible AI in defense requires human oversight, ethical safeguards, risk assessment frameworks, and clear accountability mechanisms. So, India must pursue sovereign AI with safety, trust, and accountability, strategic capability with ethical oversight, and indigenous innovation with global cooperation. India has the opportunity to become a leader in AI. We can lead responsibly, strategically, and confidently. Thank you. Jai Hind.

Thank you. Thank you, Lieutenant General Rana, CNCSC, Mr. Alkesh Kumar Sharma, former Secretary MeitY, Shri Subhash Gopinath. Ladies and gentlemen, first and foremost, I must compliment the BEL and Global team for organizing this very, very relevant topic, at least from the security agency perspective, on AI for Strategic Defense, as part of the AI Impact Summit 2026, which is absolutely imperative today. Looking at the way things are changing, operational environments for soldiers are progressively becoming complex with incorporation of advanced technologies in military operations, be it network-centric warfare, multi-domain operations, unmanned combat systems. The complexities are resulting in increased information overload, reduced reaction times, and quick decision cycles for the human operator. In order to cater to these challenges, it is the need of the hour to introduce autonomy into our military systems to augment the capabilities of the soldier. In June 2018, the Government of India published the National Strategy for AI, focusing on how India can leverage the transformative AI technologies to ensure inclusive growth in all sectors of the society, from sensing ranging from computer vision to audio processing, natural language processing, knowledge representation, machine learning, expert systems, etcetera. The AI is being classified into three variants: descriptive AI, which describes what has happened; to predictive, which is anticipating what is going to happen; and to prescriptive AI, which recommends what to do. With the emerging technological capabilities, AI is geared up to serve as a promising tool to provide enhanced situational awareness to the commanders and operators in highly contested battlefield environments, making our military systems more accurate and our military operations more swift and precise. If we look at the various domains where AI will get integrated, and some of the areas the ecosystem is working on are underwater surveillance, navigation controls, advanced material characterization, hybrid high-fidelity models in the naval systems domain, trajectory and manual planning, target identification, threat classification, aerodynamics and controls. So, I have listed out. I think for paucity of time, we will not be able to get into the details. But if you look at it, you look at armaments, combat engineering, missiles, strategic systems, electronics and communication systems, computational, cyber systems, everywhere there is an application where AI is getting seamlessly integrated. In the near future, AI will find applications in multiple operational areas, and some of the use cases I am just going to deliberate in the next few minutes. With the upcoming challenges of long-range interceptions, reduced reaction times, and saturation of airspace, AI-based decision support systems can greatly augment faster, adaptive, and intelligent decision-making across early surveillance, automated threat classification, quick and autonomous interception, and real-time battle damage assessment. Similarly, in the counter-unmanned aerial systems grid, it can be used to neutralize threats from sub-conventional aerial platforms like drones, loitering munitions, swarm drones, with the network system of AI-powered multi-layer detection and identification, dynamic resource allocation, tracking and engagement of targets through hard kill and soft kill measures. CQC operations with man-machine teaming in urban environments can be carried out using coordinated combat operations in complex, high-density environments where challenges of limited visibility, civilian presence, multi-story buildings, and ambush risk. All these can be handled by using AI-enabled advanced surveillance and integrated systems for enhancing the soldier's effectiveness, leading to very, very effective man-machine teaming. With increasing physical, psychological, and cognitive loads in highly contested environments, the future soldier and his equipment will need to be integrated with all nearby sensors and equipment. In such a scenario, AI will help in integrating the sensory information along with the commander's decision-making process through shared mental models. AI will also help in modernizing and coordinating the integrated India's EW capabilities across services with indigenous networks and real-time AI-powered electronic warfare suits like smart jammers, cognitive radios, and cyber-secure infrastructure. Weight and size optimization for performance enhancement and operations in extreme environments. Call for R&D on new-age materials where AI-based modeling and simulations will help in synthesizing the materials, providing higher accuracy, help with adaptive manufacturing, predictive quality control for high-fidelity models, and so on and so forth. In defense, reliability and robustness is a critical system requirement to prevent mission failure and unintended consequences. Ensuring that AI applications are safe, resilient to attacks, and aligned with ethical values is imperative for our defense forces. In order to address this, DRDO has identified five core trustworthy AI principles: reliability and robustness, transparency, fairness, privacy, safety, and security, which address various vulnerabilities experienced by AI systems during different lifecycle stages of development. The trustworthiness principles have been further augmented with 11 qualifying criteria, and the implementation process has been duly published as a set of two policy documents, namely, Evaluating Trustworthiness in AI Systems: The Framework Document and the Guidelines Document. As we know that AI is as good as it is trained. It heavily depends on training data, most of which is taken from the open source. With reservations, India, with its varied geography, essentially needs domestic military data infrastructure for training, validation, and testing data sets. Imported AI tools may have vulnerabilities. Also, the defense models will be classified and sensitive and need to be validated in indigenous environments. And an indigenous AI verification and validation framework will ensure that all data remains within our sovereign boundaries and Indian control, adhering to domestic policies and regulations. Indian military leadership will need interpretable AI decisions with explainability, human trust factors, and ethical considerations in local operating scenarios. The domestic framework is aligned with existing Indian policies and standards published by DGQA, SEMILAC, BIS. All these will be required to be aligned. This framework has been officially released last month and is available for the relevant stakeholders and is geared towards providing a unified approach to bringing resilience and trustworthiness in AI-powered military systems. Moving forward, we need a whole-of-nation approach for faster progression and utilization of AI-enabled solutions in the defense domain. It is our expectation that academia will establish validation and verification maturity models for each AI and explainability. And we expect industries to play a leading role in developing indigenous AI AI tools, toolkits, including model probes, adversarial testbeds, etcetera. DRDO can support in building standardized defense AI test suites for each domain and use case scenarios. And the quality assurance agencies need to instrument, need to formalize the AI verification and validation test labs and sandboxes for qualification and airworthiness. Armed forces may need to institutionalize AI safety requirements as a part of the acquisition lifecycle process. And Directorate of Standardization is expected to define and enforce the minimum mandatory verification and validation protocols for operational scenarios. We expect that NIC will be required to build infrastructure for defense-grade secure data repositories for training, validation, and testing. The roadmap towards self-reliance in defense AI lies in establishing a national AI defense strategy. And General Rana spoke about it, detailing various aspects of autonomy, decision support, threat analysis, creating dedicated structures for integrating all the stakeholders in the ecosystem, using the ethical guidelines for trust, and the domestic data infrastructure will need to be ramped up to ensure secure and federated data spanning ISR sensors, battlefield, and historical combat data. We will also need to work on labeling and annotation standards along with standardized processes which will need to be adopted across platforms and operational scenarios. A clear policy on sovereign data should be formulated and implemented at the earliest to prevent any dependency on foreign data and to ensure data residency, specifically for the defense domain. The model design must cater to explainability, adversarial robustness, and fail-safe mechanisms, and dedicated indigenous protocols for domestic combat scenarios should be adopted for verification and validation. I will just finish in another half a minute. Low-latency AI interface models, hardware accelerators, and trusted environments will be required for deployment and enablement for resilience. Cyber hardening will be required for ensuring resilience to spoofing, jamming, and cyber attacks. It is proposed that adaptive command support may be adopted, wherein AI must augment and not replace human decision-making in mission-critical tasks, and a redundant human-in-the-loop conventional override system be included for backup and failover systems. Together, I am sure we can use AI very, very effectively and efficiently in the defense domain for building strategic capability. Thank you. Jai Hind.

Thank you, Ma'am. That was truly inspiring. Today's keynotes from all three dignitaries have been extremely insightful. My sincere thanks to Lieutenant General Dinesh Singh Rana, Shri Alkesh Kumar Sharma, and Dr. Chandrika Kaushik for sharing valuable perspectives. That may play also a role in strategic deterrence. As we are talking about nuclear deterrence, we have seen in the Cold War the mutual assured destruction. Now, there is a talk about the mutual assured AI malfunction, which is already been talked about, and this may be actually a reality. ASI, that is super intelligence, having that kind of tremendous power, could also be used by a country which possesses this capability against a weaker country in the form of what we have seen in the nuclear domain, nuclear brinkmanship or nuclear blackmail. Now, when we talk about non-proliferation of these capabilities in the nuclear domain, it is possible. Presently, not very successful, but we could have some non-proliferation worldwide, so that non-state actors do not acquire them. But the same would be quite difficult because of obvious reasons when we talk about super intelligence, once it comes into being. Now, when we talk about this artificial intelligence use in strategic domain and national security, there is a lot of usage of the same. In my present appointment as Director General DIA, I have been exposed to many of open-source intelligence systems. If they combine well with the sensors of all the three services, cyber and space, and also some of the other elements like what NASA has developed, the artificial light, the light analysis software. You can really identify the hotspots emerging in your neighborhood or.

In your own area. So that kind of thing which is already in place in certain places can be further developed by us in times to come.

Now, in cyber, when we talk about AI, it inherently is more offensive in nature in cyber. We have seen the GTG 102 attack which happened in September last year. The autonomous cyber attacks by Chinese state actors against the US systems. But at the same time, the cyber capabilities can be very effectively used to identify the vulnerabilities and plug the gaps.

Now, when we talk about other usages, I think you are quite aware, right from the basic systems, from the AI software, we can, we are transitioning into command and control systems, improving our maritime domain awareness at tactical levels, electronic warfare jamming, very effectively we can do that. And what we have seen in operations, the AI can now be very effective in combating misinformation which is in vast numbers. We can, we saw in operations.

Now, the question is, what we need to do. We need to do, and as India, when we talk about the capabilities. First of all, I would suggest that we should have, we should develop an AI security doctrine, and we should use our resources where actually, as per doctrine, where it is required more. And second is, if you take an example from the Russia-Ukraine conflict, Russia, there are a lot of systems which are getting deployed without waiting from top-down capabilities. Battlefield low-cost, cheap AI-based systems which are easily scalable and are cheap are getting.

Shri Subhash Gopinath, ladies and gentlemen, first and foremost, I must compliment the Bell and Global team for organizing this very, very relevant topic, at least from the security agencies' perspective, on AI for Strategic Defense as a part of the AI Impact Summit 2026, which is absolutely imperative today, looking at the way things are changing. Operational environments for soldiers are progressively becoming complex with the incorporation of advanced technologies into military operations. Be it network-centric warfare, multi-domain operations, unmanned combat systems, the complexities are resulting in increased information overloads, reduced reaction times, and quick decision cycles for the human operator. In order to cater to these challenges, it is the need of the hour to introduce autonomy into our military systems to augment the capabilities of the soldier.

In June 2018, the Government of India published the National Strategy for AI, focusing on how India can leverage the transformative AI technologies to ensure inclusive growth in all sectors of the society, from sensing to comprehension to action. Today, AI has pervaded numerous end-user applications ranging from computer vision to audio processing, natural language processing, knowledge representation, machine learning, expert systems, etcetera. The AI is being classified into three variants: descriptive AI, which describes what has happened; predictive, which is anticipating what is going to happen; and prescriptive AI, which recommends what to do. With the emerging technological capabilities, AI is geared up to serve as a promising tool to provide enhanced situational awareness to the commanders and operators in highly contested battlefield environments, making our military systems more accurate and our military operations more swift and precise.

If we look at the various domains where AI will get integrated, and some of the areas are, the ecosystem is working on underwater surveillance, navigation controls, advanced material characterization, hybrid high-fidelity models in the naval systems domain, trajectory and manual planning, target identification, threat classification, aerodynamics, and controls. So, I have listed out. I think for paucity of time, I will not be able to get into the details. But if you look at it, you look at arms, combat engineering, missiles, strategic systems, electronics and communication systems, computational, cyber systems, everywhere there is an application where AI is getting seamlessly integrated.

In the near future, AI will find applications in multiple operational areas, and some of the use cases I am just going to deliberate in the next few minutes. With the upcoming challenges of long-range interceptions, reduced reaction times, and saturation of airspace, AI-based decision support systems can greatly augment faster, adaptive, and intelligent decision-making across early surveillance, automated threat classification, quick and autonomous interception, and real-time battle damage assessment. Similarly, in the counter-unmanned aerial systems grid, it can be used to neutralize threats from sub-conventional aerial platforms like drones, loitering munitions, swarm drones, with the network system of AI-powered multilayer detection and identification, dynamic resource allocation, tracking and engagement of targets through hard kill and soft kill measures. C4ISR operations with man-machine teaming in urban environments can be carried out using coordinated combat operations in complex, high-density environments where challenges of limited visibility, civilian presence, multi-story buildings, and ambush risks. All these can be handled by using AI-enabled advanced surveillance and integrated systems for enhancing the soldier's effectiveness, leading to very, very effective man-machine teaming. With increasing physical, psychological, and cognitive loads in highly contested environments, the future soldier and his equipment will need to be integrated with all nearby sensors and equipment. In such a scenario, AI will help in integrating the sensory information along with the commander's decision-making process through shared mental models.

AI will also help in modernizing, in coordinating the integrated India's EW capabilities across services with indigenous networks and real-time AI-powered electronic warfare suites like smart jammers, cognitive radios, and cyber-secure infrastructure. Weight and size optimization for performance enhancement and operations in extreme environments. Call for R&D on new-age materials where AI-based modeling and simulations will help in synthesizing the materials, providing higher accuracy, help with adaptive manufacturing, predictive quality control for high-fidelity models, and so and so forth.

As we know that AI is as good as it is trained. It heavily depends on training data, most of which is taken from the open source. With respect to India, with its varied geography, essentially needs domestic military data infrastructure for training, validation, and testing datasets. Imported V&V tools may have vulnerabilities. Also, the defense models will be classified and sensitive and need to be validated in indigenous environments. And an indigenous AI verification and validation framework will ensure that all data remains within our sovereign boundaries and Indian control, adhering to domestic policies and regulations. Indian military leadership will need interpretable AI decisions with explainability, human trust factors, and ethical considerations in local operating scenarios. The domestic framework is aligned with existing Indian policies and standards published by DGQA, SEMILC, BIS. All these will be required to be aligned. This framework has been officially released last month and is available for the relevant stakeholders and is geared towards providing a unified approach to bringing resilience and trustworthiness in AI-powered military systems.

The roadmap towards self-reliance in defense AI lies in establishing a national AI defense strategy. And General Rana spoke about it, detailing various aspects of autonomy, decision support, threat analysis, creating dedicated structures for integrating all the stakeholders in the ecosystem, using the ethical guidelines for AI, and the domestic data infrastructure will need to be ramped up to ensure secure and federated data spanning ISR sensors, battlefield, and historical combat data. We will also need to work on labeling and annotation standards along with standardized processes which will need to be adopted across platforms and operational scenarios. A clear policy on sovereign data should be formulated and implemented at the earliest to prevent any dependency on foreign data and to ensure data residency, specifically for the defense domain. The model designs must cater to explainability, adversarial robustness, and fail-safe mechanisms. And dedicated indigenous protocols for domestic combat scenarios should be adopted for verification and validation.

I will just finish in another half a minute. Low latency AI interface models, hardware accelerators, and trusted environments will be required for deployment and as enablers for resilience. Cyber hardening will be required for ensuring resilience to spoofing, jamming, and cyber attacks. It is proposed that an adaptive combat command support may be adopted, wherein AI must augment and not replace human decision-making in mission-critical tasks. And a redundant human-in-the-loop conventional override systems be included for backup and failover systems. Together, I am sure we can use AI very, very effectively and efficiently in the defense domain for building strategic capabilities. Thank you. Jai Hind.

Now, when we talk about non-proliferation of these capabilities in the nuclear domain, it is possible, presently not very successful, but we could have some non-proliferation worldwide, so that non-state actors don't acquire them. But the same would be quite difficult because of obvious reasons, when we talk about superintelligence, once it comes into being. Now, when we talk about this artificial intelligence use in the strategic domain and national security, there is a lot of usage of the same. In my present appointment as Director General DIA, I have been exposed to many open-source intelligence systems. If they combine well with the sensors of all the three services, cyber and space, and also some of the other elements like what NASA has developed, the artificial light, the light analysis software, you can really identify the hotspots emerging in your neighborhood or in your own area. So that kind of thing which is already in place in certain places can be further developed by us in times to come.

Now, when we talk about other usages, I think you are quite aware, right from the basic usage of logistics and preventing and maintaining your systems, from the AI software, we can, we are transitioning into command and control systems, improving our maritime domain awareness at tactical levels, electronic warfare jamming, very effectively we can do that. And what we have seen in operations, the AI can now be very effective in combating misinformation which is in vast numbers. We can, we saw in operations.

Now, the question is, what we need to do. We need to do, and as India, when we talk about the capabilities. First of all, I would suggest that we should have, we should develop an AI security doctrine, and we should use our resources where actually, as per doctrine, where it is required more. And second is, if we take an example from the Russia-Ukraine conflict. Russia, there are a lot of systems which are getting deployed without waiting from top-down capabilities. Battlefield low-cost, cheap AI-based systems which are easily scalable and are cheap are getting good results. They are increasing the kill chains of these autonomous systems and drones. So this needs to be deployed while we work on the higher systems of artificial intelligence.

Now, taking some examples back, when I was commanding a corps in Four Corps in Arunachal, where I was responsible for the security of the Line of Actual Control of Tawang and area around, I remember we had a very, we had developed a small artificial intelligence software, low-ability, but when there was an attempt by China, a foil attempt, that Yangtse, there is a place where it is disputed and only held by India, only disputed territory held by India. They, unprecedentedly, they tried their last time. There was an unprecedented attempt. While this happened for two months, through some of the, we could see that there is some issue which is building up. But finally, we could predict the time when they are coming at us. And the prediction helped us to place our resources at the right places. And we could foil the attempt of the Chinese unprecedented attempt. And thereafter, having no casualties by us, and we could place our resources and evacuation in time. So no casualties and highest, I think, unprecedented gallant award we got in the section 19 of them. So, AI prediction tools can actually help you operationally very effectively.

I would also say that we should also build agentic systems for higher fusion. And at the national level also, have all the systems of economic security and other things also deployed together so that it can help our apex level leadership to decide that what are the actions they are taking, what are the second and third-order effects of those actions, rather than looking at in isolation. The civil-military fusion is the order of the day. I think this gives responsible proactivity, and we can really enhance our capabilities there, boost our talent. And lastly, more importantly, when we become this as a superpower, which we are aspiring, and I am sure we will be on that path, we are sure we are going to become a very responsible AI power. As we are a responsible nuclear power when we acquired the nuclear capability, no first use itself shows that India is a very responsible power. And this kind of superintelligence, once we acquire, would be very effectively and responsibly used. Thank you.

Thank you, sir, for those valuable insights. We now move to our next key speaker, a distinguished civil servant who has played a foundational role in shaping India's digital and AI journey, Shri Alkesh Kumar Sharma, Senior IAS Officer of the Kerala Cadre, served as Secretary, Ministry of Electronics and Information Technology, Government of India. Shri Alkesh Kumar Sharma is currently a Member of the Public Enterprise Selection Board, Government of India, responsible for selecting top executives for Central Public Sector Enterprises. He earlier served as Secretary, Ministry of Electronics and Information Technology (MeitY), where he led major national digital initiatives, including Digital Governance, India Stack, Global Digital Public Infrastructure, Bhashini, Digital India Language Translation Platform, the National Program on AI, Semiconductor Program, Production Linked Incentive Schemes for Electronics Manufacturing, and the Digital Personal Data Protection Act, significantly shaping India's digital economy. Prior to MeitY, he served as Secretary and Additional Secretary in the Cabinet Secretariat under the Prime Minister of India, where he appraised key cabinet proposals across infrastructure, finance, agriculture, industry, and urban development, and monitored mega national projects through the PMO's Project Monitoring Group. May I have Shri Alkesh Kumar Sharma ji on the stage, please? Give him a big round of applause.

Thank you, Lieutenant General Rana, Dr. Chandrika Kaushik, Mr. Subhash, distinguished officials, my colleague in MeitY, Dr. Sanjay Bahl, distinguished friends. It is a matter of great privilege for me to be here this morning. And at the outset, I would like to congratulate both Bell and Globals for organizing this event, which is quite appropriate, part of the Global AI Summit. And last year, I remember when Prime Minister, our Prime Minister, was in the third AI Global Summit in Paris, he, under the transformative impact of AI on society, emphasized both its rapid evolution and widespread adoption. He highlighted AI expanding all in shaping politics, economics, national security, and governance. As we all know, artificial intelligence is no longer a technological advancement; it is a potential. It has the potential to drive innovation, enhance productivity, and act as a force multiplier for security in the defense. It is the strategic capability shaping economic power, geopolitical influence, and national resilience. In fact, AI is the defining capability of the 21st century, reshaping competitiveness and sovereignty. And the decisions which we take today on AI investment and governance will go a long way in defining the trajectory of nations for decades. And that is why India has taken some early lead. And it was not long ago, let me take a step back, that we started talking about AI. It was in late '21 or early '22 when we said there is something called AI and we should start doing something about it. And that is when the first AI platform, India AI, was set up. We started the first National Program on AI in 2022 as part of the Digital India initiative. And that is when, when we launched the India Stack in 2022, this Digital India initiative actually helped India in strengthening its AI policy architecture and digital governance framework. There was a push towards trusted digital public infrastructure and indigenous technology ecosystem. It gained momentum. We developed institutional frameworks for structured AI adoption across sectors, AI in health, education, agriculture. The Centers of Excellence started announcing in the budget, and we started working on that. Develop national greater emphasis placed on data governance, DPI, innovative ecosystem, and public-private partnership. And this has all, this has led to a strong base for India's AI mission, which was launched a year later, after we actually held the Global Partnership Summit in December 2023. A collaborative ecosystem between government, startups, academia, and industry was set up. A framework for safe, trusted, secure, and accountable digital transformation was set up. The foundations built during this period continue to support India's expanding AI ecosystem. Today, while developing the policy framework, we had no doubt in our mind that only a sovereign AI would help India in achieving a safe, trusted, secure, and responsible AI. As we all know, rented models will not only have biases, but data quality, security, and Indian context will be missing in those. India's sovereign AI framework will develop, control, and regulate AI systems aligned with our national interests. And it is not isolation; it is strategic autonomy combined with trusted collaboration. Today, India is witnessing an AI-driven transformation across sectors, whether it is legal, finance, healthcare, agriculture, governance, manufacturing. You talk about it. While I am now in my new role, picking up top executives for the CPsEs in India, I am finding that CPsEs are developing new AI applications which are not only increasing efficiency in operations, but they are creating higher value addition across the value chain from incremental to manifold jumps. As startups are disrupting sectors, large enterprises are embedding AI operations. As we know, our policy framework, it rests on seven important pillars. But let me come back to the AI in defense and strategic domains. India recognizes the strategic importance of AI in defense. It recognizes the need to leverage AI to enhance combat capabilities, streamline decision-making processes, and optimize resource deployment. And that's important for all our defense forces. In

This context AI offers immense force multiplier potential, enabling superior situational awareness, autonomous systems, and data defense strategies. As technology advances, the world must be aware of the latest happenings to maintain our operational superiority in defense and national security.

India's AI policy framework must ensure control over our critical data sets, reduce dependence on external AI systems, align with national doctrine and values, and protect our strategic infrastructure. Now, AI is embedded across defense systems, which Lieutenant General has talked about. But AI in defense must be controlled, auditable, and accountable. It can create havoc if it is not properly controlled. Responsible AI in defense requires human oversight, ethical safeguards, a risk assessment framework, and clear accountability mechanisms.

So, India must pursue AI with safety, trust, and accountability; strategic capability with ethical oversight; and indigenous innovation with global cooperation. India has the opportunity to become a leader in AI. We can lead responsibly, strategically, and confidently. Thank you. Jai Hind. Thank you, Sir.

We now welcome one of India's most accomplished and respected defense scientists. She is an inspiration to many, especially women aspiring to lead in science, technology, and national security. Featured by India Today as a leading women defense leader, she has made outstanding contributions to India's strategic capabilities. It is none other than Dr. Chandrika Kaushik, Distinguished Scientist and Director General, Production Coordination and Services Interaction, DRDO. Let's welcome Chandrika Kaushik, Ma'am, please.

Thank you. Lieutenant General Rana, CNCSFC, Mr. Alkesh Kumar Sharma, former Secretary, MeitY, Shri Subhash Gopinath, ladies and gentlemen. First and foremost, I must compliment the Bell and Global team for organizing this very, very relevant topic, at least from the security agencies' perspective, on AI for Strategic Defense, as part of the AI Impact Summit 2026, which is absolutely imperative today, looking at the way things are changing. Operational environments for soldiers are progressively becoming complex with the incorporation of advanced technologies into military operations. Be it network-centric warfare, multi-domain operations, unmanned combat systems, the complexities are resulting in increased overloads, reduced reaction times, and quick decision-making for the human operator. In order to cater to these challenges, it is the need of the hour to introduce autonomy into our military systems to augment the capabilities of the soldier.

In June 2018, the Government of India published the National Strategy for AI, focusing on how India can leverage the transformative AI technologies to ensure inclusive growth in all sectors of society, from sensing to comprehension to action. Today, AI has pervaded numerous end-user applications, ranging from computer vision to audio processing, natural language processing, knowledge representation, machine learning, expert systems, etc. AI is being classified into three variants: descriptive AI, which describes what has happened; predictive, which is anticipating what is going to happen; and prescriptive AI, which recommends what to do. With the emerging technological capabilities, AI is geared up to serve as a promising tool to provide enhanced situational awareness to the commanders and operators in highly contested battlefield environments, making our military systems more accurate and our military operations more swift and precise.

If we look at the various domains where AI will get integrated, some of the areas our ecosystem is working on are underwater surveillance, navigation controls, advanced material characterization, hybrid high-fidelity models in the naval systems domain, trajectory and manual planning, target identification, threat classification, aerodynamics, and controls. So, I have listed out. I think for paucity of time, I will not be able to get into the details. But if you look at it, you look at arms, combat engineering, missiles, strategic systems, electronics and communication systems, computational cyber systems, everywhere there is an application where AI is getting seamlessly integrated.

With increasing physical, psychological, and cognitive loads in highly contested environments, the future soldier and his equipment will need to be integrated with all nearby sensors and equipment. In such a scenario, AI will help in integrating the sensory information along with the commander's decision-making process through shared mental models. AI will also help in modernizing and coordinating the integrated India's EW capabilities across services with indigenous networks and real-time AI-powered electronic warfare suites like smart jammers, cognitive radios, and cyber-secure infrastructure, weight and size optimization for performance enhancement, and operations in extreme environments. Call for R&D on new-age materials where AI-based modeling and simulations will help in synthesizing these materials, providing higher accuracy, help with adaptive manufacturing, predictive quality control for high-fidelity models, and so on and so forth.

In defense, reliability and robustness are critical system requirements to prevent mission failure and unintended consequences. Ensuring that AI applications are safe, resilient to attacks, and aligned with ethical values is imperative for our defense forces. In order to address this, DRDO has identified five core trustworthy AI principles: reliability and robustness, transparency, fairness, privacy, safety, and security, which address various vulnerabilities experienced by AI systems during different life cycle stages of development. The trustworthiness principles have been further augmented with 11 qualifying criteria, and the implementation process has been duly published as a set of two policy documents, namely "Evaluating Trustworthiness in AI Systems: The Framework Document" and "The Guidelines Document."

As we know that AI is as good as it is trained. It heavily depends on training data, most of which is taken from the open source. With reservations, India, with its varied geography, essentially needs domestic military data infrastructure for training, validation, and testing data sets. Imported V&V tools may have vulnerabilities. Also, the defense models will be classified and sensitive and need to be validated in an indigenous environment. An indigenous AI verification and validation framework will ensure that all data remains within our sovereign boundaries and Indian control, adhering to domestic policies and regulations. Indian military leadership will need interpretable AI decisions with explainability, human trust factors, and ethical considerations in local operating scenarios. The domestic framework is aligned with existing Indian policies and standards published by DGQA, SEMILC, BIS. All these will be required to be aligned. This framework has been officially released last month and is available for the relevant stakeholders and is geared towards providing a unified approach to bringing resilience and trustworthiness in AI-powered military systems.

The roadmap towards self-reliance in defense AI lies in establishing a national AI defense strategy. And General Rana spoke about it, detailing various aspects of autonomy, decision support, threat analysis, creating dedicated structures for integrating all the stakeholders in the ecosystem, using the ethical guidelines for safety, and the domestic data infrastructure will need to be ramped up to ensure secure and federated data spanning ISR sensors, battlefield, and historical combat data. We will also need to work on labeling and annotation standards along with standardized processes which will need to be adopted across platforms and operational scenarios. A clear policy on sovereign data should be formulated and implemented at the earliest to prevent any dependency on foreign data and to ensure data residency specifically for the defense domain. The model design must cater to explainability, adversarial robustness, and fail-safe mechanisms. And dedicated indigenous protocols for domestic combat scenarios should be adopted for verification and validation.

Now, when we talk about non-proliferation of these capabilities in the nuclear domain is possible, presently not very successful, but we could have some non-proliferation worldwide, so that non-state actors don't acquire them. But the same would be quite difficult because of obvious reasons. When we talk about superintelligence, once it comes into being. Now, when we talk about artificial intelligence use in the strategic domain and national security, there is a lot of usage of the same. In my present appointment as Director General, DIA, I have been exposed to many open-source intelligence systems. If they combine well with the sensors of all the three services, cyber and space, and also some of the other elements like what NASA has developed, the artificial light, the light analysis software, you can really identify the hotspots emerging in your neighborhood or in your own area. So that kind of thing, which is already in place in certain places, can be further developed by us in times to come.

Now, in cyber, when we talk about AI, inherently it is more offensive in nature. In cyber, we have seen the GTG 102 attack which happened in September last year, the autonomous cyber attacks by Chinese state actors against the US systems. But at the same time, the cyber capabilities can be very effectively used to identify the vulnerabilities and plug the gaps. Now, when we talk about other usages, I think you are quite aware, right from the basic usage of logistics and preventing and maintaining your systems from the AI software. We can, we are transitioning into command and control systems, improving our maritime domain awareness at tactical levels, electronic warfare jamming, very effectively we can do that. And what we have seen in operations, AI can now be very effective in combating misinformation, which is in vast numbers. We can, we saw in operations Sindoor.

Now, the question is, what we need to do. We need to do, and as India, when we talk about the capabilities. First of all, I would suggest that we should have, we should develop an AI security doctrine. And we should use our resources where actually, as per doctrine, where it is required more. And second is, if you take an example from the Russia-Ukraine conflict, Russia, there are a lot of systems which are getting deployed without waiting from top-down capabilities. Battlefield, low-cost AI-based systems which are easily scalable and are cheap are giving good results. They are increasing the kill chains of these autonomous systems and drones. So this needs to be deployed while we work on the higher systems of artificial intelligence.

Now, taking some examples back, when I was commanding a corps in IV Corps in Arunachal, where I was responsible for the security of the Line of Actual Control of Tawang and areas around, I remember we had a very, we had developed a small artificial intelligence software, low-ability, but when there was an attempt by China, a foil attempt, that Yangtse, there is a place where it is disputed and only held by India, only disputed territory held by India. They, unprecedentedly, they tried their last time. There was an unprecedented attempt. While this happened for two months, through some of the systems, we could see that there is some issue which is building up. But finally, we could predict the time when they are coming at us. And the prediction helped us to place our resources at the right places. And we could foil the attempt of the Chinese unprecedented attempt. And thereafter, having no casualties by us. And we could place our resources and evacuation in time. So no casualties and highest, I think, unprecedented gallant award we got in the section 19 of them. So artificial intelligence prediction tools can actually help you operate any very effectively.

I would also say that we should also build agentic systems for higher fusion. And at the national level also, have all the systems of economic security and other things also deployed together. So that it can help our apex level leadership to decide that what are the actions they are taking, what are the second and third-order effects of those actions, rather than looking at them in isolation. The civil-military fusion is the order of the day. I think this gives our responsible activity, and we can really enhance our capabilities there, boost our talent. And lastly, more importantly, when we become this, as a superpower which we are aspiring, and I am sure we will be on that path, I am sure we are going to become a very responsible AI power. As we are a responsible nuclear power, when we acquired the nuclear capability, our "no first use" itself shows that India is a very responsible power. And this kind of superintelligence, once we acquire, would be very effectively and responsibly used. Thank you.

Thank you, Sir, for those valuable insights. We now move to our next keynote speaker, a distinguished civil servant who has played a foundational role in shaping India's digital and AI journey. Shri Alkesh Kumar Sharma, a senior IAS officer of the Kerala cadre, served as Secretary, Ministry of Electronics and Information Technology, Government of India. Shri Alkesh Kumar Sharma is currently a Member of the Public Enterprises Selection Board, Government of India, responsible for selecting top executives for Central Public Sector Enterprises. He earlier served as Secretary, Ministry of Electronics and Information Technology, where he led major national digital initiatives, including Digital Governance, India Stack, Global Digital Public Infrastructure, Bhashini, Digital India Language Translation Platform, the National Program on AI, Semiconductor Program, Production Linked Incentive Schemes for Electronics Manufacturing, and the Digital Personal Data Protection Act, significantly shaping India's digital economy. Prior to that, he served as Secretary and Additional Secretary in the Cabinet Secretariat under the Prime Minister of India, where he appraised key cabinet proposals across infrastructure, finance, agriculture, industry, and urban development, and monitored mega national projects through the PMO's Project Monitoring Group. May I have Shri Alkesh Kumar Sharma Ji on the stage, please?

Yes, thank you. Lieutenant General Rana, Dr. Chandrika Kaushik, Mr. Subhash, distinguished officials, my colleagues in MeitY, Dr. Sanjay Bahal, distinguished friends. It is a matter of great privilege for me to be here this morning. And at the outset, I would like to congratulate both Bell and Globals for organizing this event, which is quite an appropriate part of the Global AI Summit. And last year, I remember when Prime Minister, our Prime Minister, was at the third AI Global Summit in Paris, he underscored the transformative impact of AI on society, emphasizing both its rapidly evolving nature and widespread adoption. He highlighted AI expanding all spheres, shaping politics, economics, national security, and governance. As we all know, artificial intelligence is no longer a technological advancement; it is a potential, it has the potential to drive innovation, enhance productivity, and act as a force multiplier for security. In defense, it is the strategic capability, shaping economic power, geopolitical influence, and national resilience. AI is the defining capability of the 21st century, reshaping competitiveness and sovereignty. And the decisions which we take today on AI investment and governance will go a long way in defining the trajectory of nations for decades. And that is why India has taken some early lead. And it was not long ago, let me take a step back, that we started talking about AI. It was in late 2021 or early 2022 when we said there is something called AI and we should start doing something about it. And that is when the first AI platform, India, was set up. We started the first National Program on AI in 2022 as part of the Digital India initiative. And that is when, when we launched the India Stack in 2022, this Digital India initiative actually helped India in strengthening its AI policy architecture and digital governance framework. There was a push towards trusted digital public infrastructure and indigenous technology ecosystem. It gained momentum. We developed institutional frameworks for structured AI adoption across sectors: AI in health, education, agriculture. The Centers of Excellence system started. Announcing the budget, and we started working on that. Developing national greater emphasis was placed on data governance, DPI, innovative ecosystem, and public-private partnership. And this has all, this has led to a strong base for India's AI mission, which was launched a year later, after we actually held the Global Partnership Summit in December 2023. A collaborative ecosystem between government, startups, academia, and industry was set up. A framework for safe, trusted, secure, and accountable digital transformation was set up. The foundations built during this period continue to support India's expanding AI ecosystem.

Today, while developing the policy framework, we had no doubt in our minds that only a sovereign AI would help India in achieving a safe, trusted, secure, and responsible AI. As we all know, rented models will not only have biases, but data quality, security, and Indian context will be missing in those. India's sovereign AI framework will develop, control, and regulate AI systems aligned with our national interest. And it is not isolation; it is strategic autonomy combined with a trusted collaboration. Today, India is witnessing AI-driven transformation across sectors, whether it is legal, finance, healthcare, agriculture, governance, manufacturing. You talk about it. While I am now in my new role, picking up top executives for the CPSC in India, I am finding that CPs are developing new AI applications which are not only increasing efficiency in operations, but they are creating higher value addition across the value chain, from incremental to manifold gains. As startups are disrupting traditional sectors, large enterprises are embedding AI operations. As we know, our policy framework rests on seven important pillars. But let me come back to the AI in defense and strategic importance. India recognizes the strategic importance of AI in defense. It recognizes the need to leverage AI to enhance combat capabilities, streamline decision-making processes, and optimize resource deployment. And that's important for all our defense forces. In this context, AI offers immense force multiplier potential, enabling superior situational awareness, autonomous systems, and data-driven strategies. As technology advances, the world must be aware of the latest happenings.

To maintain our operational superiority in defense and national security contexts, India's AI policy framework must ensure control over our critical datasets, reduce dependence on external AI systems, alignment with national doctrine and values, and protection of our strategic infrastructure. Now, AI is embedded across defense systems, which Lieutenant General has talked about. But AI in defense must be controlled, auditable, and accountable. It can create havoc if it is not properly controlled. Responsible AI in defense requires human oversight, ethical safeguards, risk assessment frameworks, and clear accountability mechanisms. So, India must pursue AI with safety, trust, and accountability; strategic capability with ethical oversight; and indigenous innovation with global cooperation. India has the opportunity to become a leader in AI. We can lead responsibly, strategically, and confidently. Thank you. Jai Hind. Thank you, sir.

Lieutenant General Rana, CNCSFC, Mr. Alkesh Kumar Sharma, former Secretary, MeitY, Shri Subhash Gopinath, ladies and gentlemen. First and foremost, I must compliment the Bell and Global team for organizing this very, very relevant topic, at least from the security agencies' perspective, on AI for Strategic Defense, as part of the AI Impact Summit 2026, which is absolutely imperative today. Looking at the way things are changing, operational environments for soldiers are progressively becoming complex with the incorporation of advanced technologies into military operations. Be it network-centric warfare, multi-domain operations, unmanned combat systems, these complexities are resulting in increased information overloads, reduced reaction times, and quick decision cycles for the human operator. In order to cater to these challenges, it is the need of the hour to introduce autonomy into our military systems to augment the capabilities of the soldier.

In June 2018, the Government of India published the National Strategy for AI, focusing on how India can leverage the transformative AI technologies to ensure inclusive growth in all sectors of the society. From sensing, ranging from computer vision to audio processing, natural language processing, knowledge representation, machine learning, expert systems, etcetera. The AI is being classified into three variants: descriptive AI, which describes what has happened; predictive, which is anticipating what is going to happen; and prescriptive AI, which recommends what to do. With the emerging technological capabilities, AI is geared up to serve as a promising tool to provide enhanced situational awareness to the commanders and operators in highly contested battlefield environments, making our military systems more accurate and our military operations more swift and precise.

If we look at the various domains where AI will get integrated, and some of these areas are: ecosystem is working on underwater surveillance, navigation controls, advanced material characterization, hybrid high-fidelity models in the naval systems domain, trajectory and maneuver planning, target identification, threat classification, aerodynamics, and controls. So, I have listed out. I think for paucity of time, I will not be able to get into the details. But if you look at it, you look at arms, combat engineering, missiles, strategic systems, electronics and communication systems, computational cyber systems, everywhere there is an application where AI is getting seamlessly integrated. In the near future, AI will find applications in multiple operational areas, and some of the use cases I am just going to deliberate in the next few minutes. With the upcoming challenges of long-range interceptions, reduced reaction times, and saturation of airspace, AI-based decision support systems can greatly augment faster, adaptive, and intelligent decision-making across early surveillance, automated threat classification, quick and autonomous interception, and real-time battle damage assessment. Similarly, in the counter-unmanned aerial systems grid, it can be used to neutralize threats from sub-conventional aerial platforms like drones, loitering munitions, swarm drones, with the network system of AI-powered multi-layer detection and identification, dynamic resource allocation, tracking, and engagement of targets through hard kill and soft kill measures. C4ISR operations with man-machine teaming in urban environments can be carried out using coordinated combat operations in complex, high-density environments where challenges of limited visibility, civilian presence, multi-storied buildings, and ambush risk. All these can be handled by using AI-enabled advanced surveillance and integrated systems for enhancing the soldier's effectiveness, leading to very, very effective man-machine teaming. With increasing physical, psychological, and cognitive loads in highly contested environments, the future soldier and his equipment will need to be integrated with all nearby sensors and equipment. In such a scenario, AI will help in integrating the sensory information along with the commander's decision-making process through shared mental models.

AI will also help in modernizing and coordinating the integrated India's EW capabilities across services with indigenous networks and real-time AI-powered electronic warfare suites like smart jammers, cognitive radios, and cyber-secure infrastructure. Weight and size optimization for performance enhancement and operations in extreme environments call for R&D on new-age materials where AI-based modeling and simulations will help in synthesizing these materials, providing higher accuracy, help with adaptive manufacturing, predictive quality control for high-fidelity models, and so on and so forth. In defense, reliability and robustness is a critical system requirement to prevent mission failure and unintended consequences. Ensuring that AI applications are safe, resilient to attacks, and aligned with ethical values is imperative for our defense forces. In order to address this, DRDO has identified five core trustworthy AI principles: reliability and robustness, transparency, fairness, privacy, safety, and security, which address various vulnerabilities experienced by AI systems during different life cycle stages of development. The trustworthiness principles have been further augmented with 11 qualifying criteria, and the implementation process has been duly published as a set of two policy documents, namely, "Evaluating Trustworthiness in AI Systems: The Framework Document" and "The Guidelines Document."

As we know that AI is as good as it is trained, it heavily depends on training data, most of which is taken from the open source. With respect to India, with its varied geography, essentially needs domestic military data infrastructure for training, validation, and testing. Datasets imported via NV tools may have vulnerabilities. Also, the defense models will be classified and sensitive and need to be validated in indigenous environments. An indigenous AI verification and validation framework will ensure that all data remains within our sovereign boundaries and Indian control, adhering to domestic policies and regulations. Indian military leadership will need interpretable AI decisions with explainability, human trust factors, and ethical considerations in local operating scenarios. The domestic framework is aligned with existing Indian policies and standards published by DGQA, CEL, BIS. All these will be required to be aligned. This framework has been officially released last month and is available for the relevant stakeholders and is geared towards providing a unified approach to bringing resilience and trustworthiness in AI-powered military systems.

Moving forward, we need a whole-of-nation approach for faster progression and utilization of AI-enabled solutions in the defense domain. It is our expectation that academia will establish validation and verification maturity models for each AI and explainability. And we expect industries to play a leading role in developing indigenous AI V&V toolkits, toolchains, including model probes, adversarial testbeds, etcetera. DRDO can support in building standardized defense AI test suites for each domain and use case scenarios. And the quality assurance agencies need to instrument, need to formalize the AI V&V test labs and sandboxes for qualification and airworthiness. Armed forces may need to institutionalize AI safety requirements as a part of the acquisition life cycle process. And the Directorate of Standardization is expected to define and enforce the minimum mandatory V&V protocols for operational scenarios. We expect that NIC will be required to build infrastructure for defense-grade secure data repositories for training, validation, and testing. The roadmap towards self-reliance in defense AI lies in establishing a national AI defense strategy. And General Rana spoke about it, detailing various aspects of autonomy, decision support, threat analysis, creating dedicated structures for integrating all the stakeholders in the ecosystem, using the ethical guidelines for. And the domestic data infrastructure will need to be ramped up to ensure secure and federated data spanning ISR sensors, battlefield, and historical combat data. We will also need to work on labeling and annotation standards along with standardized processes which will need to be adopted across platforms and operational scenarios. A clear policy on sovereign data should be formulated and implemented at the earliest to prevent any dependency on foreign data and to ensure data residency, specifically for the defense domain. The model design must cater to explainability, adversarial robustness, and fail-safe mechanisms, and dedicated indigenous protocols for domestic combat scenarios should be adopted for verification and validation.

I will just finish in another half minute. Low latency AI interface models, hardware accelerators, and trusted environments will be required for deployment and as enablement for resilience. Cyber hardening will be required for ensuring resilience to spoofing, jamming, and cyber attacks. It is proposed that adaptive combat command support may be adopted, wherein AI must augment and not replace human decision-making in mission-critical tasks, and a redundant human-in-the-loop conventional override system be included for backup and failover systems. Together, I am sure we can use AI very, very effectively and efficiently in the defense domain for building strategic capability. Thank you. Jai Hind.

Now, when we talk about artificial intelligence in the strategic domain and national security, there is a lot of usage of the same. In my present appointment as Director General, DIA, I have been exposed to many open-source intelligence systems. If they combine well with the sensors of all the three services, cyber and space, and also some of the other elements like what NASA has developed, the artificial light, the light analysis software, you can really identify the hotspots emerging in your neighborhood or in your own area. So that kind of thing, which is already in place in certain places, can be further developed by us in times to come. Now, in cyber, when we talk about AI, inherently it is more offensive in nature. In cyber, we have seen the GTG 102 attack which happened in September last year, the autonomous cyber attacks by Chinese state actors against the US systems. But at the same time, the cyber capabilities can be very effectively used to identify the vulnerabilities and plug the gaps.

Now, when we talk about other usages, I think you are quite aware, right from the basic systems, from the AI software, we are transitioning into command and control systems, improving our maritime domain awareness at tactical levels, electronic warfare jamming very effectively we can do that. And what we have seen in Operation Sindhu, the AI can now be very effective in combating misinformation, which is in vast numbers. We saw in Operation Sindhu. Now, the question is, what we need to do? We need to do, and as India, when we talk about the capabilities. First of all, I would suggest that we should have, we should develop an AI security doctrine, and we should use our resources where actually, as per doctrine, where it is required more. And second is, if you take an example from the Russia-Ukraine conflict, in Russia, there are a lot of systems which are getting deployed without waiting from top-down capabilities. Battlefield low-cost AI-based systems which are easily scalable and are cheap are getting deployed.

Shri Subhash Gopinath, ladies and gentlemen. First and foremost, I must compliment the Bell and Global team for organizing this very, very relevant topic, at least from the security agencies' perspective, on AI for Strategic Defense, as part of the AI Impact Summit 2026, which is absolutely imperative today. Looking at the way things are changing, operational environments for soldiers are progressively becoming complex with the incorporation of advanced technologies into military operations. Be it network-centric warfare, multi-domain operations, unmanned combat systems, these complexities are resulting in increased information overloads, reduced reaction times, and quick decision cycles for the human operator. In order to cater to these challenges, it is the need of the hour to introduce autonomy into our military systems to augment the capabilities of the soldier.

In June 2018, the Government of India published the National Strategy for AI, focusing on how India can leverage the transformative AI technologies to ensure inclusive growth in all sectors of the society. From sensing to comprehension to action. Today, AI has pervaded numerous end-user applications, ranging from computer vision to audio processing, natural language processing, knowledge representation, machine learning, expert systems, etcetera. The AI is being classified into three variants: descriptive AI, which describes what has happened; predictive, which is anticipating what is going to happen; and prescriptive AI, which recommends what to do. With the emerging technological capabilities, AI is geared up to serve as a promising tool to provide enhanced situational awareness to the commanders and operators in highly contested battlefield environments, making our military systems more accurate and our military operations more swift and precise.

Moving forward, we need a whole-of-nation approach for faster progression and utilization of AI-enabled solutions in the defense domain. It is our expectation that academia will establish validation and verification maturity models for each AI and explainability. And we expect industries to play a leading role in developing indigenous AI V&V toolkits, toolchains, including model probes, adversarial testbeds, etcetera. DRDO can support in building standardized defense AI test suites for each domain and use case scenarios. And the quality assurance agencies need to instrument, need to formalize the AI V&V test labs and sandboxes for qualification and airworthiness. Armed forces may need to institutionalize AI safety requirements as a part of the acquisition life cycle process. And the Directorate of Standardization is expected to define and enforce the minimum mandatory V&V protocols for operational scenarios. We expect that NIC will be required to build infrastructure for defense-grade secure data repositories for training, validation, and testing. The roadmap towards self-reliance in defense AI lies in establishing a national AI defense strategy. And General Rana spoke about it, detailing various aspects of autonomy, decision support, threat analysis, creating dedicated structures for integrating all the stakeholders in the ecosystem, using the ethical guidelines for. And the domestic data infrastructure will need to be ramped up to ensure secure and federated data spanning ISR sensors, battlefield, and historical combat data. We will also need to work on labeling and annotation standards along with standardized processes which will need to be adopted across platforms and operational scenarios. A clear policy on sovereign data should be formulated and implemented at the earliest to prevent any dependency on foreign data and to ensure data residency, specifically for the defense domain. The model design must cater to explainability, adversarial robustness, and fail-safe mechanisms, and dedicated indigenous protocols for domestic combat scenarios should be adopted for verification and validation.

Decision making in mission critical tasks and a redundant human in the loop conventional override systems be included for backup and failover systems. Together, I am sure we can use AI very, very effectively and efficiently in the defense domain for building strategic capability. Thank you. Jai Hind. Thank you, ma'am. That was truly inspiring. Today's keynotes from all three dignitaries have been extremely insightful. My sincere thanks to Lieutenant General Dinesh Singh Rana, Shri Alkesh Kumar Sharma, and Dr. Chandrika Kaushik for sharing valuable perspectives that may also play a role in strategic deterrence. As we are talking about nuclear deterrence, we have seen in the Cold War the mutual assured destruction. Now, there is a talk about the mutual assured AI malfunction, which is already been talked about, and this may be actually a reality. ASI, that is superintelligence, having that kind of tremendous power, could also be used by a country which possesses possesses this capability against a weaker country in the form of what we have seen in the nuclear domain, nuclear brinkmanship or nuclear blackmail.

Now, when we talk about non-proliferation of the capabilities in the nuclear domain, is it possible? Presently, not very successful. But we could have some non-proliferation worldwide, so that non-state actors do not acquire them. But the same would be quite difficult because of obvious reasons when we talk about superintelligence, once it comes into being. Now, when we talk about this artificial intelligence choice in the strategic domain and national security, there is a lot of usage of the same. In my present appointment as Director General DIA, I have been exposed to many open-source intelligence systems. If they combine well with the sensors of all the three services, cyber and space, and also some of the other elements like what NASA has developed, the artificial light, the light analysis software, you can really identify the hotspots emerging in your neighborhood or in your own area. So, that kind of thing, which is already in place in certain places, can be further developed by us in times to come.

Now, in cyber, when we talk about AI, inherently it is more offensive in nature in cyber. We have seen the GTG 102 attack which happened in September last year, the autonomous cyber attacks by Chinese state actors against the US systems. But at the same time, the cyber capabilities can be very effectively used to identify the vulnerabilities and plug the gaps. Now, when we talk about other usages, I think you are quite aware, right from the basic usage of logistics and preventing and maintaining your systems from the AI software, we can, we are transitioning to command and control systems, improving our maritime domain awareness at tactical levels, electronic warfare jamming very effectively, we can do that. And what we have seen in operations, the AI can now be very effective in combating misinformation, which is in vast numbers. We saw in operations Sindoor.

Now, the question is, what we need to do? We need to do, and as India, when we talk about the capabilities. First of all, I would suggest that we should have, we should develop an AI security doctrine, and we should use our resources where actually, as per the doctrine, where it is required more. And second is, if we take an example from the Russia-Ukraine conflict, Russia, there are a lot of systems which are getting deployed without waiting from top-down capabilities. Battlefield, low-cost AI-based systems which are easily scalable and are cheap, are getting giving good results. They are increasing the kill chains of these autonomous systems and drones. So, this needs to be deployed, while we work on the higher systems of artificial intelligence.

Now, taking some examples back, when I was commanding a corps in Four Corps, in Arunachal, where I was responsible for the security of the Line of Actual Control of Tawang and area around. I remember we had a very, we had developed a small artificial intelligence software, low capability. But when there was an attempt by China, a false attempt, that Yangse, there is a place where it is disputed and only held by India, only disputed territory held by India. They, unprecedentedly, they tried their last time there. An attempt. While this happened for two months, through some of the systems, we could see that there is some issue which is building up. But finally, we could predict the time when they are coming at us. And the prediction helped us to place our resources at the right places. And we could foil the attempt of the Chinese, unprecedented attempt. And thereafter, having no casualties by us. And we could place our resources and evacuation in time. So, no casualties and highest, I think, unprecedented guarantee award we got in Section 19 of them. So, artificial intelligence prediction tools can actually help you operate very effectively. I would also say that we should also build agentic systems for higher fusion and at the national level also have all the systems of economic security and other things also deployed together. So that it can help our apex level leadership to decide what are the actions they are taking, what are the second and third order effects of those actions, rather than looking at them in isolation. The civil-military fusion is the order of the day. I think this gives our responsible activity and we can really enhance our capabilities there. Boost our talent. And lastly, more importantly, when we become this, as a superpower which we are aspiring, and I am sure we will be on that path. I am sure we are going to become a very responsible AI power, as we are a responsible nuclear power. When we acquired nuclear capability, our no-first-use itself shows that India is a very responsible power. And this kind of superintelligence, once we acquire, would be very effectively and responsibly used. Thank you.

Thank you, sir, for those valuable insights. We now move to our next keynote speaker, a distinguished civil servant who has played a foundational role in shaping India's digital and AI journey. Shri Alkesh Kumar Sharma, Senior IAS officer of the Kerala cadre, served as Secretary, Ministry of Electronics and Information Technology, Government of India. Shri Alkesh Kumar Sharma is currently a member of the Public Enterprises Selection Board, Government of India, responsible for selecting top executives for Central Public Sector Enterprises. He earlier served as Secretary, Ministry of Electronics and Information Technology (MeitY), where he led major national digital initiatives, including Digital Governance, India's Tech Stack, Global Digital Public Infrastructure, Bhashini, Digital India Language Translation Platform, the National Program on AI, Semiconductor Program, Production Linked Incentive schemes for electronics manufacturing, and the Digital Personal Data Protection Act, significantly shaping India's digital economy. Prior to MeitY, he served as Secretary and Additional Secretary in the Cabinet Secretariat under the Prime Minister of India, where he approved key cabinet proposals across infrastructure, finance, agriculture, industry, and urban development, and monitored mega national projects through the PMO's Project Monitoring Group. May I have Shri Alkesh Kumar Sharma ji on the stage, please?

Yes. Thank you, Lieutenant General Rana, Dr. Kaushik, Mr. Subhash, distinguished army officials, my colleague in MeitY, Dr. Sanjay Bahl, distinguished friends. It is a matter of great privilege for me to be here this morning. And at the outset, I would like to congratulate both Bael and Globals for organizing this event, which is quite appropriate, part of the Global AI Summit. And last year, I remember when Prime Minister, our Prime Minister, was in the third AI Global Summit in Paris, he underscored the transformative impact of AI on society, emphasizing both its rapid evolution and widespread adoption. He highlighted AI expanding all in shaping politics, economics, national security, and governance. As we all know, artificial intelligence is no longer a technological advancement; it is a potential. It has the potential to drive innovation, enhance productivity, and act as a force multiplier for security. In defense, it is the strategic capability shaping economic power, geopolitical influence, and national resilience. In fact, AI is between the defining capability of the 21st century, reshaping competitiveness and sovereignty. And the decisions which we take today on AI investment and governance will go a long way in defining the trajectory of nations for decades. And that is why India has taken some early lead. And it was not long ago, let me take a step back, that we started talking about AI. It was in late '21 or early '22 when we said there is something called AI and we should start doing something about it. And that is when the first AI platform India was set up. We started the first National Program on AI in 2022 as part of the Digital India initiative. And that is when, when we launched the India Stack in 2022, this Digital India initiative actually helped India in strengthening its AI policy architecture and digital governance framework. There is a push towards trusted digital public infrastructure and indigenous technology ecosystem. It gained momentum. We developed institutional frameworks for structured AI adoption across sectors, AI in health, education, agriculture. The Centers of Excellence started announced in the budget, and we started working on that. Develop national greater emphasis placed on data governance, DPI, innovative ecosystem, and public-private partnership. And this has all, this has led to a strong base for India's AI Mission, which was launched a year later after we actually held the Global Partnership Summit in December 2023. A collaborative ecosystem between government, startups, academia, and industry was set up. A framework for safe, trusted, secure, and accountable digital transformation was set up. The foundations built during this period continue to support India's expanding AI ecosystem today. While developing the policy framework, we had no doubt in our mind that only a sovereign AI would help India in achieving a safe, trusted, secure, and responsible AI. As we all know, rented models will not only have biases, but data quality, security, and Indian context will be missing in those. India's sovereign AI framework will develop, control, and regulate AI systems aligned with our national interest. And it is not in isolation; it is strategic autonomy combined with trusted collaboration. Today, India is witnessing AI-driven transformation across sectors, whether it is legal, finance, healthcare, agriculture, governance, manufacturing, you talk about it. And while I am now in my new role, picking up top executives for the CPSC in India, I am finding that CPs are developing new AI applications which are not only increasing efficiency in operations, but they are creating higher value addition across the value chain from incremental to manifold gains. As startups are disrupting traditional sectors, large enterprises are embedding AI operations. As we know, our policy framework, its rests on seven important pillars. But let me come back to the AI in defense and strategic importance. India recognizes the strategic importance of AI in defense. It recognizes the need to leverage AI to enhance combat capabilities, streamline decision-making processes, and optimize resource deployment, and that's important for all our defense forces. In this context, AI offers immense force multiplier potential, enabling superior situational awareness, autonomous systems, and data defense strategies. As technology, the world must be aware of the latest what is happening in to maintain our operational superiority. In defense and national security context, India, the AI policy framework must ensure control over our critical datasets, reduce dependence on external AI systems, alignment with national doctrine and values, and protection of our strategic infrastructure. Now, AI embedded across defense systems, which Lieutenant General has talked about, but AI in defense must be controlled, auditable, and accountable. It can create havoc if it is not properly controlled. Responsible AI in defense requires human oversight, ethical safeguards, risk assessment frameworks, and clear accountability mechanisms. So, India must pursue sovereign AI with safety, trust, and accountability, strategic capability with ethical oversight, and indigenous innovation with global cooperation. India has the opportunity to become a leader in AI. We can lead responsibly, strategically, and confidently. Thank you. Jai Hind.

Thank you, sir. We now welcome one of India's most accomplished and respected defense scientists. She is an inspiration to many, especially women aspiring to lead in science, technology, and national security. Featured by India Today as a leading woman defense leader, she has made outstanding contributions to India's strategic capabilities. It is none other than Dr. Chandrika Kaushik, Distinguished Scientist and Director General, Production Coordination and Services Interaction, DRDO. Let's welcome Chandrika Kaushik, ma'am, please.

Thank you. Presentation, Lieutenant General Rana, CNC, SFC, Mr. Alkesh Kumar Sharma, former Secretary, MeitY, Mr. Subhash Gopinath, ladies and gentlemen. First and foremost, I must compliment the Bael and Global team for organizing this very, very relevant topic, at least from the security agencies' perspective, on AI for Strategic Defense, as part of the AI Impact Summit 2026, which is absolutely imperative today, looking at the way things are changing. Operational environments for soldiers are progressively becoming complex with the incorporation of advanced technologies into military operations, be it network-centric warfare, multi-domain operations, unmanned combat systems. The complexities are resulting in increased overloads, reduced reaction times, and quick decision cycles for the human operator. In order to cater to these challenges, it is the need of the hour to introduce autonomy into our military systems to augment the capabilities of the soldier. In June 2018, the Government of India published the National Strategy for AI, focusing on how India can leverage the transformative AI technologies to ensure inclusive growth in all sectors of the society, from sensing to comprehension to action. Today, AI has pervaded numerous end-user applications ranging from computer vision to audio processing, natural language processing, knowledge representation, machine learning, expert systems, etcetera. The AI is being classified into three variants: descriptive AI, which describes what has happened; to predictive, which is anticipating what is going to happen; and to prescriptive AI, which recommends what to do. With the emerging technological capabilities, AI is geared up to serve as a promising tool to provide enhanced situational awareness to the commanders and operators in highly contested battlefield environments, making our military systems more accurate and our military operations more swift and precise. If we look at the various domains where AI will get integrated, and some of these areas are, the ecosystem is working on underwater surveillance, navigation controls, advanced material characterization, hybrid high-fidelity models in the naval systems domain, trajectory and maneuver planning, target identification, threat classification, aerodynamics and controls. So, I have listed out, I think, for paucity of time, I will not be able to get into the details. But if you look at it, you look at armaments, combat engineering, missiles, strategic systems, electronics and communication systems, computational cyber systems, everywhere there is an application where AI is getting seamlessly integrated. In the near future, AI will find applications in multiple operational areas, and some of the use cases, I am just going to deliberate in the next few minutes. With the upcoming challenges of long-range interceptions, reduced reaction times, and saturation of airspace, AI-based decision support systems can greatly augment faster, adaptive, and intelligent decision-making across early surveillance, automated threat classification, quick and autonomous interception, and real-time battle damage assessment. Similarly, in the counter-unmanned aerial systems grid, it can be used to neutralize threats from sub-conventional aerial platforms like drones, loitering munitions, swarm drones, with the network system of AI-powered multi-layer detection and identification, dynamic resource allocation, tracking and engagement of targets through hard kill and soft kill measures. C4ISR operations with man-machine teaming in urban environments can be carried out using coordinated combat operations in complex, high-density environments where challenges of limited visibility, civilian presence, multi-storied buildings, and ambush risks, all these can be handled by using AI-enabled advanced surveillance and integrated systems for enhancing the soldier's effectiveness, leading to very, very effective man-machine teaming. With increasing physical, psychological, and cognitive loads in highly contested environments, the future soldier and his equipment will need to be integrated with all nearby sensors and equipment. In such a scenario, AI will help in integrating the sensory information along with the commander's decision-making process through shared mental models. AI will also help in modernizing and coordinating the integrated India's EW capabilities across services with indigenous networks and real-time AI-powered electronic warfare suits like smart jammers, cognitive radios, and cyber-secure infrastructure. Weight and size optimization for performance enhancement and operations in extreme environments call for R&D on new-age materials where AI-based modeling and simulations will help in synthesizing these materials, providing higher accuracy, help with adaptive manufacturing, predictive quality control for high-fidelity models, and so on and so forth. In defense, reliability and robustness is a critical system requirement to prevent mission failure and unintended consequences. Ensuring that AI applications are safe, resilient to attacks, and aligned with ethical values is imperative for our defense forces. In order to address this, DRDO has identified five core trustworthy AI principles: reliability and robustness, transparency, fairness, privacy, safety, and security, which address various vulnerabilities experienced by AI systems during different lifecycle stages of development. The trustworthiness principles have been further augmented with 11 qualifying criteria, and the implementation process has been duly published as a set of two policy documents, namely, "Evaluating Trustworthiness in AI Systems: The Framework Document" and "The Guidelines Document." As we know that AI is as good as it is trained. It heavily depends on training data, most of which is taken from the open source. With respect to India, with its varied geography, essentially needs domestic military data infrastructure for training, validation, and testing data sets. Imported V&V tools may have vulnerabilities. Also, the defense models will be classified and sensitive and need to be validated in indigenous environments. And an indigenous AI verification and validation framework will ensure that all data remains within our sovereign boundaries and Indian control, adhering to domestic policies and regulations. Indian military leadership will need interpretable AI decisions with explainability, human trust factors, and ethical considerations in local operating scenarios. The domestic framework is aligned with existing Indian policies and standards published by DGQA, CILAC, BIS. All these will be required to be aligned. This framework has been officially released last month and is available for the relevant stakeholders and is geared towards providing a unified approach to bringing resilience and trustworthiness in AI-powered military systems. Moving forward, we need a whole-of-nation approach for faster progression and utilization of AI-enabled solutions in the defense domain. It is our expectation that academia will establish validation and verification maturity models for each AI and explainability, and we expect industries to play a leading role in developing indigenous AI V&V toolkits, toolchains, including model probes, adversarial testbeds, etcetera. DRDO can support in building standardized defense AI test suits for each domain and use case scenarios, and the quality assurance agencies need to instrument, need to formalize the AI V&V test labs and sandboxes for qualification and airworthiness. Armed forces may need to institutionalize AI safety requirements as part of the acquisition lifecycle process, and the Directorate of Standardization is expected to define and enforce the minimum mandatory V&V protocols for operational scenarios. We expect that NIC will be required to build infrastructure for defense-grade secure data repositories for training, validation, and testing. The roadmap towards self-reliance in defense AI lies in establishing a national AI defense strategy. And General Rana spoke about it, detailing various aspects of autonomy, decision support, threat analysis, creating dedicated structures for integrating all the stakeholders in the ecosystem, using the ethical guidelines for standards, and the domestic data infrastructure will need to be ramped up to ensure secure and federated data spanning ISR sensors, battlefield, and historical combat data. We will also need to work on labeling and annotation standards along with standardized processes which will need to be adopted across platforms and operational scenarios. A clear policy on sovereign data should be formulated and implemented at the earliest to prevent any dependency on foreign data and to ensure data residency, specifically for the defense domain. The model design must cater to explainability, adversarial robustness, and fail-safe mechanisms, and dedicated indigenous protocols for domestic combat scenarios should be adopted for verification and validation. I will just finish in another half a minute. Low latency AI interface models, hardware accelerators, and trusted environments will be required for deployment and as enablers for resilience. Cyber hardening will be required for ensuring resilience to spoofing, jamming, and cyber attacks. It is proposed that adaptive command support may be adopted, wherein AI must augment and not replace human decision-making in mission critical tasks, and a redundant human in the loop conventional override systems be included for backup and failover systems. Together, I am sure we can use AI very, very effectively and efficiently in the defense.

Domain for building strategic capability. Thank you. Jai Hind. Thank you, ma'am. That was truly inspiring. Today's keynotes from all three dignitaries have been extremely insightful. My sincere thanks to Lieutenant General Dinesh Singh Rana, Shri Alkesh Kumar Sharma, and Dr. Chandrika Kaushik for sharing valuable perspectives that may also play a role in strategic deterrence.

As we are talking about nuclear deterrence, we have seen in the Cold War the concept of Mutual Assured Destruction. Now, there is talk about the Mutual Assured AI Malfunction, which has already been discussed. This may actually be a reality. AI, that is, superintelligence, having a kind of tremendous power, could also be used by a country which possesses this capability against a weaker country, in the form of what we have seen in the nuclear domain: nuclear brinkmanship or nuclear blackmail.

Now, when we talk about non-proliferation of these capabilities in the nuclear domain, it is possible. Presently, it is not very successful, but we could have some non-proliferation worldwide, so that non-state actors don't acquire them. But the same would be quite difficult because of obvious reasons when we talk about superintelligence, once it comes into being.

Now, when we talk about artificial intelligence use in the strategic domain and national security, there is a lot of usage of the same. In my previous appointment as Director General DIA, I have been exposed to many open-source intelligence systems. If they combine well with the sensors of all the three services, cyber and space, and also some of the other elements like what NASA has developed, the artificial light analysis software, you can really identify the hotspots emerging in your neighborhood or in your own area. So, that kind of thing, which is already in place in certain places, can be further developed by us in times to come.

Now, in cyber, when we talk about AI, it inherently is more offensive in nature. In cyber, we have seen the GTG 102 attack which happened in September last year, the autonomous cyber attacks by Chinese state actors against US systems. But at the same time, the cyber capabilities can be very effectively used to identify the vulnerabilities and plug the gaps.

Now, when we talk about other usages, I think you are quite aware, right from the basic usage of logistics and preventing and maintaining your systems from the AI software. We can, we are transitioning to command and control systems, improving our maritime domain awareness at tactical levels, electronic warfare jamming very effectively, we can do that. And what we have seen in operations, AI can now be very effective in combating misinformation, which is in vast numbers, we saw in operations.

And second is, if we take an example from the Russia-Ukraine conflict. Russia, there are a lot of systems which are getting deployed without waiting for top-down capabilities. Battlefield, low-cost AI-based systems, which are easily scalable and are cheap, are giving good results. They are increasing the kill chains of these autonomous systems and drones. So, this needs to be deployed, while we work on the higher systems of artificial intelligence.

Now, taking some examples back, when I was commanding a corps in four corps, in Arunachal, where I was responsible for the security of the Line of Actual Control of Tawang and areas around. I remember we had a very, we had developed a small artificial intelligence software, low capability. But when there was an attempt by China, a failed attempt, that Yangtse, there is a place where it is disputed and only held by India, only disputed territory held by India. They, unprecedentedly, they tried there last time. An unprecedented attempt. While this happened for two months, through some of the systems, we could see that there is some issue which is building up. But finally, we could predict the time when they are coming at us. And the prediction helped us to place our resources at the right places. And we could foil the attempt of the Chinese, an unprecedented attempt. And thereafter, having no casualties by us. And we could place our resources and evacuation in time. So, no casualties and highest, I think, unprecedented guarantee award we got in Section 19 of them. So, artificial intelligence prediction tools can actually help you operationally very effectively.

I would also say that we should also build agentic systems for higher fusion and at the national level also have all the systems of economic security and other things also deployed together. So that it can help our apex level leadership to decide what are the actions they are taking, what are the second and third-order effects of those actions, rather than looking at them in isolation. The civil-military fusion is the order of the day. I think this gives our responsible activity, and we can really enhance our capabilities there. Boost our talent. And lastly, more importantly, when we become this, as a superpower which we are aspiring to be, and I am sure we will be on that path. I am sure we are going to become a very responsible AI power, as we are a responsible nuclear power. When we acquired the nuclear capability, our "no first use" itself shows that India is a very responsible power. And this kind of superintelligence, once we acquire, would be very effectively and responsibly used. Thank you.

Thank you, sir, for those valuable insights. We now move to our next keynote speaker, a distinguished civil servant who has played a foundational role in shaping India's digital and AI journey. Shri Alkesh Kumar Sharma, a senior IAS officer of the Kerala cadre, serves as Secretary, Ministry of Electronics and Information Technology, Government of India. Shri Alkesh Kumar Sharma is currently a member of the Public Enterprises Selection Board, Government of India, responsible for selecting top executives for Central Public Sector Enterprises. He earlier served as Secretary, Ministry of Electronics and Information Technology, where he led major national digital initiatives, including Digital Governance, India Stack, Global Digital Public Infrastructure, Bhashini, Digital India Language Transition Platform, the National Program on AI, Semiconductor Program, Production Linked Incentive schemes for electronics manufacturing, and the Digital Personal Data Protection Act, significantly shaping India's digital economy. Prior to this, he served as Secretary and Additional Secretary in the Cabinet Secretariat under the Prime Minister of India, where he appraised key cabinet proposals across infrastructure, finance, agriculture, industry, and urban development, and monitored mega national projects through the PMO's Project Monitoring Group. May I have Shri Alkesh Kumar Sharma ji on stage, please?

Yes. Thank you, Lieutenant General Rana, CNC, SFC, Mr. Alkesh Kumar Sharma, former Secretary, MeitY, Dr. Sanjay Bahal, distinguished friends. It is a matter of great privilege for me to be here this morning. And at the outset, I would like to congratulate both BEL and Globals for organizing this event, which is quite appropriate as part of the Global AI Summit. And last year, I remember when our Prime Minister was at the third AI Global Summit in Paris, he underscored the transformative impact of AI on society, emphasizing both its rapid evolution and widespread adoption. He highlighted AI's expanding role in shaping politics, economics, national security, and governance. As we all know, artificial intelligence is no longer a technological advancement; it has the potential to drive innovation, enhance productivity, and act as a force multiplier for security in the defense. It is the strategic capability shaping economic power, geopolitical influence, and national resilience. In fact, AI is the defining capability of the 21st century, reshaping competitiveness and sovereignty. And the decisions which we take today on AI investment and governance will go a long way in defining the trajectory of nations for decades. And that is why India has taken some early lead. And it was not long ago, let me take a step back, that we started talking about AI. It was in late 2021 or early 2022 when we said there is something called AI and we should start doing something about it. And that is when the first AI platform, India, was set up. We started the first National Program on AI in 2022 as part of the Digital India initiative. And that is when, when we launched the India Stack in 2022, this Digital India initiative actually helped India in strengthening its AI policy architecture and digital governance framework. There is a push towards trusted digital public infrastructure and indigenous technology ecosystem. It gained momentum. We developed institutional frameworks for structured AI adoption across sectors, AI in health, education, agriculture, with the Center of Excellence systems. We started announcing the budget, and we started working on that. We developed national guidelines, placed emphasis on data governance, DPI, innovative ecosystem, and public-private partnership. And this has all, this has led to a strong base for India's AI mission, which was launched a year later after we actually held the Global Partnership Summit in December 2023. A collaborative ecosystem between government, startups, academia, and industry was set up. A framework for safe, trusted, secure, and accountable digital transformation was set up. The foundations built during this period continue to support India's expanding AI ecosystem today. While developing the policy framework, we had no doubt in our minds that only a sovereign AI would help India in achieving a safe, trusted, secure, and responsible AI. As we all know, rented models will not only have biases, but data quality, security, and Indian context will be missing in those. India's sovereign AI framework will develop, control, and regulate AI systems aligned with our national interests. And it is not in isolation; it is strategic autonomy combined with trusted collaboration. Today, India is witnessing AI-driven transformation across sectors, whether it is legal, finance, healthcare, agriculture, governance, manufacturing. You talk about it. And while I am now in my new role, picking up top executives for the CPsEs in India, I am finding that CPsEs are developing new AI applications, which are not only increasing efficiency in operations, but they are creating higher value addition across the value chain, from incremental to manifold gains. As the startups are disrupting traditional sectors, large enterprises are embedding AI in operations. As we know, our policy framework rests on seven important pillars. But let me come back to AI in defense and strategic capabilities. India recognizes the strategic importance of AI in defense. It recognizes the need to leverage AI to enhance combat capabilities, streamline decision-making processes, and optimize resource deployment. And that's important for all our defense forces. In this context, AI offers immense force multiplier potential, enabling superior situational awareness, autonomous systems, and data defense strategies. As technology world, we must be aware of the latest what is happening to maintain our operational superiority. In defense and national security context, India's AI policy framework must ensure control over critical datasets, reduce dependence on external AI systems, alignment with national doctrine and values, and protection of our strategic infrastructure. Now, AI embedded across defense systems, which Lieutenant General has talked about. But AI in defense must be controlled, auditable, and accountable. It can create havoc if it is not properly controlled. Responsible AI in defense requires human oversight, ethical safeguards, risk assessment frameworks, and clear accountability mechanisms. So, India must pursue sovereign AI with safety, trust, and accountability, strategic capability with ethical oversight, and indigenous innovation with global cooperation. India has the opportunity to become a leader in AI. We can lead responsibly, strategically, and confidently. Thank you. Jai Hind.

Thank you. Presentation, Lieutenant General Rana, CNC, SFC, Mr. Alkesh Kumar Sharma, former Secretary, MeitY, Mr. Subhash Gopinath, ladies and gentlemen. First and foremost, I must compliment the BEL and Global team for organizing this very, very relevant topic, at least from the security agencies' perspective, on AI for Strategic Defense, as part of the AI Impact Summit 2026, which is absolutely imperative today. Looking at the way things are changing, operational environments for soldiers are progressively becoming complex with the incorporation of advanced technologies into military operations. Be it network-centric warfare, multi-domain operations, unmanned combat systems, the complexities are resulting in increased information overloads, reduced reaction times, and quick decision cycles for the human operator. In order to cater to these challenges, it is the need of the hour to introduce autonomy into our military systems to augment the capabilities of the soldier. In June 2018, the Government of India published the National Strategy for AI, focusing on how India can leverage the transformative AI technologies to ensure inclusive growth in all sectors of the society, from sensing ranging from computer vision to audio processing, natural language processing, knowledge representation, machine learning, expert systems, etcetera. The AI is being classified into three variants: descriptive AI, which describes what has happened; predictive, which is anticipating what is going to happen; and prescriptive AI, which recommends what to do. With the emerging technological capabilities, AI is geared up to serve as a promising tool to provide enhanced situational awareness to the commanders and operators in highly contested battlefield environments, making our military systems more accurate and our military operations more swift and precise. If we look at the various domains where AI will get integrated, and some of the areas the ecosystem is working on are underwater surveillance, navigation controls, advanced material characterization, hybrid high-fidelity models in the naval systems domain, trajectory and maneuver planning, target identification, threat classification, aerodynamics, and controls. So, I have listed out. I think for paucity of time, I will not be able to get into the details. But if you look at it, you look at arms, combat engineering, missiles, strategic systems, electronics and communication systems, computational cyber systems, everywhere there is an application where AI is getting seamlessly integrated. In the near future, AI will find applications in multiple operational areas. And some of the use cases I am just going to deliberate in the next few minutes. With the upcoming challenges of long-range interceptions, reduced reaction times, and saturation of airspace, AI-based decision support systems can greatly augment faster, adaptive, and intelligent decision-making across early surveillance, automated threat classification, quick and autonomous interception, and real-time battle damage assessment. Similarly, in the counter-unmanned aerial systems grid, it can be used to neutralize threats from sub-conventional aerial platforms like drones, loitering munitions, swarm drones, with the network system of AI-powered multi-layer detection and identification, dynamic resource allocation, tracking and engagement of targets through hard kill and soft kill measures. C4ISR operations with man-machine teaming in urban environments can be carried out using coordinated combat operations in complex, high-density environments where challenges of limited visibility, civilian presence, multi-storied buildings, and ambush risks. All these can be handled by using AI-enabled advanced surveillance and integrated systems for enhancing the soldier's effectiveness, leading to very, very effective man-machine teaming. With increasing physical, psychological, and cognitive loads in highly contested environments, the future soldier and his equipment will need to be integrated with all nearby sensors and equipment. In such a scenario, AI will help in integrating the sensory information along with the commander's decision-making process through shared mental models. AI will also help in modernizing and coordinating the integrated India's EW capabilities across services with indigenous networks and real-time AI-powered electronic warfare suits like smart jammers, cognitive radio, and cyber-secure infrastructure. Weight and size optimization for performance enhancement and operations in extreme environments call for RAD or new-age materials where AI-based modeling and simulations will help in synthesizing these materials, providing higher accuracy, help with adaptive manufacturing, predictive quality control for high-fidelity models, and so on and so forth. In defense, reliability and robustness is a critical system requirement to prevent mission failure and unintended consequences. Ensuring that AI applications are safe, resilient to attacks, and aligned with ethical values is imperative for our defense forces. In order to address this, DRDO has identified five core trustworthy AI principles: reliability and robustness, transparency, fairness, privacy, safety, and security, which address various vulnerabilities experienced by AI systems during different lifecycle stages of development. The trustworthiness principles have been further augmented with 11 qualifying criteria, and the implementation process has been duly published as a set of two policy documents, namely, "Evaluating Trustworthiness in AI Systems: The Framework Document" and "The Guidelines Document." As we know that AI is as good as it is trained. It heavily depends on training data, most of which is taken from the open source. With respect to India, with its varied geography, essentially needs domestic military data infrastructure for training, validation, and testing data sets. Imported V&V tools may have vulnerabilities also. The defense models will be classified and sensitive and need to be validated in indigenous environments. And an indigenous AI verification and validation framework will ensure that all data remains within our sovereign boundaries and Indian control, adhering to domestic policies and regulations. Indian military leadership will need interpretable AI decisions with explainability, human trust factors, and ethical considerations in local operating scenarios. The domestic framework is aligned with existing Indian policies and standards published by DGQA, CEMILAC, BIS. All these will be required to be aligned. This framework has been officially released last month and is available for the relevant stakeholders and is geared towards providing a unified approach to bringing resilience and trustworthiness in AI-powered military systems. Moving forward, we need a whole-of-nation approach for faster progression and utilization of AI-enabled solutions in the defense domain. It is our expectation that academia will establish validation and verification maturity models for each AI and explainability. And we expect industry to play a leading role in developing indigenous AI V&V toolkits, toolchains, including model probes, adversarial test beds, etcetera. DRDO can support in building standardized defense AI test suits for each domain and use case scenarios. And the quality assurance agencies need to formalize the AI V&V test labs and sandboxes for qualification and airworthiness. Armed forces may need to institutionalize AI safety requirements as a part of the acquisition lifecycle process. And the Directorate of Standardization is expected to define and enforce the minimum mandatory V&V protocols for operational scenarios. We expect that NIC will be required to build infrastructure for defense-grade secure data repositories for training, validation, and testing. The roadmap towards self-reliance in defense AI lies in establishing a national AI defense strategy. And General Rana spoke about it, detailing various aspects of autonomy, decision support, threat analysis, creating dedicated structures for integrating all the stakeholders in the ecosystem, using the ethical guidelines for standards, and the domestic data infrastructure will need to be ramped up to ensure secure and federated data spanning ISR sensors, battlefield, and historical combat data. We will also need to work on labeling and annotation standards along with standardized processes which will need to be adopted across platforms and operational scenarios. A clear policy on sovereign data should be formulated and implemented at the earliest to prevent any dependency on foreign data and to ensure data residency, specifically for the defense domain. The model design must cater to explainability, adversarial robustness, and fail-safe mechanisms, and dedicated indigenous protocols for domestic combat scenarios should be adopted for verification and validation. I will just finish in another half a minute. Low latency AI interface models, hardware accelerators, and trusted environments will be required for deployment and as enablers for resilience. Cyber hardening will be required for ensuring resilience to spoofing, jamming, and cyber attacks. It is proposed that adaptive command support may be adopted, wherein AI must augment and not replace human decision-making in mission-critical tasks. And a redundant human-in-the-loop conventional override system be included for backup and failover systems. Together, I am sure we can use AI very, very effectively and efficiently in the defense domain for building strategic capabilities. Thank you. Jai Hind.

Play also plays a role in strategic deterrence. As we are talking about nuclear deterrence, we have seen in the Cold War the mutual assured destruction. Now, there is talk about the mutual assured AI malfunction, which has already been talked about, and this may actually be a reality. ASI, that is, superintelligence, having a kind of tremendous power, could also be used by a country which possesses this capability against a weaker country, in the form of what we have seen in the nuclear domain: nuclear brinkmanship or nuclear blackmail.

Now, when we talk about non-proliferation of these capabilities in the nuclear domain, it is possible. Presently, it is not very successful, but we could have some non-proliferation worldwide, so that non-state actors don't acquire them. But the same would be quite difficult, because of obvious reasons, when we talk about superintelligence, once it comes into being.

Now, when we talk about artificial intelligence use in the strategic domain and national security, there are a lot of uses of the same. In my present appointment as Director General DIA, I have been exposed to many open-source intelligence systems. If they combine well with the sensors of all the three services, cyber and space, and also some of the other elements like what NASA has developed, the artificial light, the light analysis software, you can really identify the hotspots emerging in your neighborhood or in your own area. So, that kind of thing, which is already in place in certain places, can be further developed by us in times to come.

Now, in cyber, when we talk about AI, it is inherently more offensive in nature. In cyber, we have seen the GTG 102 attack, which happened in September last year, the autonomous cyber attacks by Chinese state actors against the US systems. But at the same time, the cyber capabilities can be very effectively used to identify the vulnerabilities and plug the gaps.

Now, when we talk about other uses, I think you are quite aware, right from the basic systems, from the AI software, we are transitioning to command and control systems, improving our maritime domain awareness at tactical levels, electronic warfare jamming, very effectively we can do that. And what we have seen in Operation Sundew, the AI can now be very effective in combating misinformation, which is in vast numbers. We saw in Operation Sundew.

Now, the question is, what we need to do. We need to do, and as India, when we talk about these capabilities. First of all, I would suggest that we should have, we should develop an AI security doctrine, and we should use our resources where actually, as per the doctrine, where it is required more. And second is, if you take an example from the Russia-Ukraine conflict, Russia, there are a lot of systems which are getting deployed without waiting for top-down capabilities. Battlefield, low-cost AI-based systems, which are easily scalable and are cheap, are getting...

Shri Subhash Gopinath, ladies and gentlemen, first and foremost, I must compliment the Bell and Global team for organizing this very, very relevant topic, at least from the security agency perspective, on AI for Strategic Defense, as part of the AI Impact Summit 2026, which is absolutely imperative today. Looking at the way things are changing, operational environments for soldiers are progressively becoming complex with the incorporation of advanced technologies in military operations. Be it network-centric warfare, multi-domain operations, unmanned combat systems, these complexities are resulting in increased information overloads, reduced reaction times, and quick decision cycles for the human operator. In order to cater to these challenges, it is the need of the hour to introduce autonomy into our military systems to augment the capabilities of the soldier.

In June 2018, the Government of India published the National Strategy for AI, focusing on how India can leverage the transformative AI technologies to ensure inclusive growth in all sectors of society, from sensing to comprehension to action. Today, AI has pervaded numerous end-user applications, ranging from computer vision to audio processing, natural language processing, knowledge representation, machine learning, expert systems, etcetera. AI is being classified into three variants: descriptive AI, which describes what has happened; predictive, which is anticipating what is going to happen; and prescriptive AI, which recommends what to do. With the emerging technological capabilities, AI is geared up to serve as a promising tool to provide enhanced situational awareness to the commanders and operators in highly contested battlefield environments, making our military systems more accurate and our military operations more swift and precise.

In the near future, AI will find applications in multiple operational areas. And some of the use cases, I am just going to deliberate in the next few minutes. With the upcoming challenges of long-range interceptions, reduced reaction times, and saturation of airspace, AI-based decision support systems can greatly augment faster, adaptive, and intelligent decision-making across early surveillance, automated threat classification, quick and autonomous interception, and real-time battle damage assessment. Similarly, in the counter-unmanned aerial systems grid, it can be used to neutralize threats from sub-conventional aerial platforms like drones, loitering munitions, swarm drones, with the network system of AI-powered multi-layer detection and identification, dynamic resource allocation, tracking and engagement of targets through hard kill and soft kill measures. C4ISR operations with man-machine teaming in urban environments can be carried out using coordinated combat operations in complex, high-density environments where challenges of limited visibility, civilian presence, multi-story buildings, and ambush risk, all these can be handled by using AI-enabled advanced surveillance and integrated systems for enhancing the soldier's effectiveness, leading to very, very effective man-machine teaming.

In defense, reliability and robustness are critical system requirements to prevent mission failure and unintended consequences. Ensuring that AI applications are safe, resilient to attacks, and aligned with ethical values is imperative for our defense forces. In order to address this, DRDO has identified five core trustworthy AI principles: reliability and robustness, transparency, fairness, privacy, safety, and security, which address various vulnerabilities experienced by AI systems during different life cycle stages of development. The trustworthiness principles have been further augmented with 11 qualifying criteria, and the implementation process has been duly published as a set of two policy documents, namely, "Evaluating Trustworthiness in AI Systems: The Framework Document" and "The Guidelines Document."

As we know that AI is as good as it is trained. It heavily depends on training data, most of which is taken from the open source with reservations. India, with its varied geography, essentially needs domestic military data infrastructure for training, validation, and testing data sets. Imported V&V tools may have vulnerabilities. Also, the defense models will be classified and sensitive and need to be validated in indigenous environments. And indigenous AI verification and validation framework will ensure that all data remains within our sovereign boundaries and Indian control, adhering to domestic policies and regulations. Indian military leadership will need interpretable AI decisions with explainability, human trust factors, and ethical considerations in local operating scenarios. The domestic framework is aligned with existing Indian policies and standards published by DGQA, SEMILC, BIS. All these will be required to be alive. This framework has been officially released last month and is available for the relevant stakeholders and is geared towards providing a unified approach to bringing resilience and trustworthiness in AI-powered military systems.

The roadmap towards self-reliance in defense AI lies in establishing a national AI defense strategy. And General Rana spoke about it, detailing various aspects of autonomy, decision support, threat analysis, creating dedicated structures for integrating all the stakeholders in the ecosystem, using ethical guidelines, and the domestic data infrastructure will need to be ramped up to ensure secure and federated data spanning ISR sensors, battlefield, and historical combat data. We will also need to work on labeling and annotation standards along with standardized processes which will need to be adopted across platforms and operational scenarios. A clear policy on sovereign data should be formulated and implemented at the earliest to prevent any dependency on foreign data and to ensure data residency, specifically for the defense domain. The model design must cater to explainability, adversarial robustness, and fail-safe mechanisms, and dedicated indigenous protocols for domestic combat scenarios should be adopted for verification and validation.

I will just finish in another half minute. Low latency AI interface models, hardware accelerators, and trusted environments will be required for deployment and as enablers for resilience. Cyber hardening will be required for ensuring resilience to spoofing, jamming, and cyber attacks. It is proposed that adaptive combat command support may be adopted, wherein AI must augment and not replace human decision-making in mission-critical tasks, and a redundant human-in-the-loop conventional override system be included for backup and failover systems. Together, I am sure we can use AI very, very effectively and efficiently in the defense domain for building strategic capability. Thank you. Jai Hind.

That may also play a role in strategic deterrence. As we are talking about nuclear deterrence, we have seen in the Cold War the mutual assured destruction. Now, there is talk about the mutual assured AI malfunction, which has already been talked about, and this may actually be a reality. ASI, that is, superintelligence, having a kind of tremendous power, could also be used by a country which possesses this capability against a weaker country, in the form of what we have seen in the nuclear domain: nuclear brinkmanship or nuclear blackmail.

Now, when we talk about non-proliferation of these capabilities in the nuclear domain, it is possible. Presently, it is not very successful, but we could have some non-proliferation worldwide, so that non-state actors don't acquire them. But the same would be quite difficult, because of obvious reasons, when we talk about superintelligence, once it comes into being.

Now, when we talk about this artificial intelligence use in the strategic domain and national security, there are a lot of uses of the same. In my present appointment as Director General DIA, I have been exposed to many open-source intelligence systems. If they combine well with the sensors of all the three services, cyber and space, and also some of the other elements like what NASA has developed, the artificial light, the light analysis software, you can really identify the hotspots emerging in your neighborhood or in your own area. So, that kind of thing, which is already in place in certain places, can be further developed by us in times to come.

Now, when we talk about other uses, I think you are quite aware, right from the basic usage of logistics and preventing and maintaining your systems, from the AI software, we are transitioning into command and control systems, improving our maritime domain awareness at tactical levels, electronic warfare jamming, very effectively we can do that. And what we have seen in Operation Sundew, the AI can now be very effective in combating misinformation, which is in vast numbers. We saw in Operation Sundew.

Now, the question is, what we need to do. We need to do, and as India, when we talk about these capabilities. First of all, I would suggest that we should have, we should develop an AI security doctrine, and we should use our resources where actually, as per the doctrine, where it is required more. And second is, if we take an example from the Russia-Ukraine conflict, Russia, there are a lot of systems which are getting deployed without waiting for top-down capabilities. Battlefield, low-cost AI-based systems, which are easily scalable and are cheap, are getting good results. They are increasing the kill chains of these autonomous systems and drones.

So, this needs to be deployed, even while we work on the higher systems of artificial intelligence. Now, taking some examples back, when I was commanding a corps in Four Corps, in Arunachal, where I was responsible for the security of the Line of Actual Control of Tawang and the area around. I remember we had a very, we had developed a small artificial intelligence software, low capability, but when there was an attempt by China, a failed attempt, that Yangtse, there is a place where it is disputed and only held by India, only disputed territory held by India. They, unprecedentedly, they tried their last time there, an attempt. An unprecedented attempt. While this happened for two months, through some of the systems, we could see that there is some issue which is building up. But finally, we could get a prediction of the time when they are coming at us. And the prediction helped us to place our resources at the right places, and we could foil the attempt of the Chinese, an unprecedented attempt. And thereafter, having no casualties by us, and we could place our resources and evacuation in time. So, no casualties and the highest, I think, unprecedented award we got in Section 19 of them. So, artificial intelligence prediction tools can actually help you operate very effectively.

I would also say that we should also build agentic systems for higher fusion and at the national level also have all the systems of economic security and other things also deployed together, so that it can help our apex-level leadership to decide that what are the actions they are taking, what are the second and third-order effects of those actions, rather than looking at it in isolation. The civil-military fusion is the order of the day. I think this gives a responsible proactivity, and we can really enhance our capabilities there, boost our talent. And lastly, more importantly, when we become this as a superpower, which we are aspiring to be, and I am sure we will be on that path, I am sure we are going to become a very responsible AI power. As we are a responsible nuclear power, then we acquire the nuclear capability. No first use itself shows that India is a very responsible power. And this kind of superintelligence, once we acquire, would be very effectively and responsibly used.

Thank you, sir, for those valuable insights. We now move to our next keynote speaker, a distinguished civil servant who has played a foundational role in shaping India's digital and AI journey, Shri Alkesh Kumar Sharma, Senior IAS Officer of the Kerala Cadre, served as Secretary, Ministry of Electronics and Information Technology, Government of India. Shri Alkesh Kumar Sharma is currently a Member of the Public Enterprises Selection Board, Government of India, responsible for selecting top executives for Central Public Sector Enterprises. He earlier served as Secretary, Ministry of Electronics and Information Technology (MeitY), where he led major national digital initiatives, including Digital Governance, India Stack, Global Digital Public Infrastructure, Bhashini, Digital India Language Translation Platform, the National Program on AI, Semiconductor Program, Production Linked Incentive Schemes for Electronics Manufacturing, and the Digital Personal Data Protection Act, significantly shaping India's digital economy. Prior to MeitY, he served as Secretary and Additional Secretary in the Cabinet Secretariat under the Prime Minister of India, where he apprised key cabinet proposals across infrastructure, finance, agriculture, industry, and urban development, and monitored mega national projects through the PMO's Project Monitoring Group. May I have Shri Alkesh Kumar Sharma ji on the stage, please?

Yes, thank you. Lieutenant General Rana, Dr. Chandrika Kaushik, Mr. Subhash, distinguished officials, my colleagues in MeitY, Dr. Sanjay Bahl, distinguished friends. It is a matter of great privilege for me to be here this morning. And at the outset, I would like to congratulate both Bell and Globals for organizing this event, which is quite appropriate, part of the Global AI Summit. And last year, I remember when Prime Minister, our Prime Minister, was in the third AI Global Summit in Paris, he underscored the transformative impact of AI on society, emphasizing both its rapid evolution and widespread adoption. He highlighted AI expanding all in shaping politics, economics, national security, and governance. As we all know, artificial intelligence is no longer a technological advancement; it is a potential. It has the potential to drive innovation, enhance productivity, and act as a force multiplier for security in the defense. It is the strategic capability shaping economic power, geopolitical influence, and national resilience. In fact, AI is the defining capability of the 21st century, reshaping competitiveness and sovereignty. And the decisions which we take today on AI investment and governance will go a long way in defining the trajectory of nations for decades. And that is why India has taken some early lead. And it was not long ago. Let me take a step back. That we started talking about AI. It was in late 2021 or early 2022 when we said there is something called AI and we should start doing something about it. And that is when the first AI platform, India, was set up. We started the first National Program on AI in 2022 as part of the Digital India initiative. And that is when, when we launched the India Stack in 2022, this Digital India initiative actually helped India in strengthening its AI policy architecture and digital governance framework. There is a push towards trusted digital public infrastructure and indigenous technology ecosystem. It gained momentum. We developed institutional frameworks for structured AI adoption across sectors: AI in health, education, agriculture. The Centers of Excellence system started, announcing the budget, and we started working on that. Develop national greater emphasis placed on data governance, DPI, innovative ecosystem, and public-private partnership. And this has all, this has led to a strong base for India's AI mission, which was launched a year later, after we actually held the Global Partnership Summit in December 2023. A collaborative ecosystem between government, startups, academia, and industry was set up. A framework for safe, trusted, secure, and accountable digital transformation was set up. The foundations built during this period continue to support India as it expands.

The AI ecosystem today. While developing the policy framework, we had no doubt in our minds that only a sovereign AI would help India in achieving a safe, trusted, secure, and responsible AI. As we all know, rented models will not only have biases, but data quality, security, and Indian context will be missing in those. India's sovereign AI framework will develop, control, and regulate AI systems aligned with our national interests, and it is not isolation; it is strategic autonomy combined with trusted collaboration.

Today, India is witnessing AI-driven transformation across sectors. Whether it is legal, finance, healthcare, agriculture, governance, manufacturing, you name it. While I am now in my new role picking up top executives for the CPSC's in India, I am finding that CPSC's are developing new AI applications, which are not only increasing efficiency in operations, but they are creating higher value addition across the value chain from incremental to manifold gains. As the startups are dipping into the sector, large enterprises are embedding AI operations.

As we know, our policy framework rests on seven important pillars. But let me come back to the AI in defense and strategic issues. India recognizes the strategic importance of AI in defense. It recognizes a need to leverage AI to enhance combat capabilities, streamline decision-making processes, and optimize resource deployment, and that's important for all our defense forces. In this context, AI offers immense force multiplier potential, enabling superior situational awareness, autonomous systems, and data defense strategies. As technology, the world must be aware of the latest happenings to maintain our operational superiority in defense and national security.

The context of India's AI policy framework must ensure control over critical datasets, reduce dependence on external AI systems, alignment with national doctrine and values, and protection of our strategic infrastructure. Now, AI is embedded across defense systems, which Lieutenant General has talked about. But AI in defense must be controlled, auditable, and accountable. It can create havoc if it is not properly controlled. Responsible AI in defense requires human oversight, ethical safeguards, risk assessment frameworks, and clear accountability mechanisms. So, India must pursue sovereign AI with safety, trust, and accountability, strategic capability with ethical oversight, and indigenous innovation with global cooperation. India has the opportunity to become a leader in AI. We can lead responsibly, strategically, and confidently. Thank you. Jai Hind. Thank you, sir.

We now welcome one of India's most accomplished and respected defense scientists. She is an inspiration to many, especially women aspiring to lead in science, technology, and national security. Featured by India Today as a leading woman defense leader, she has made outstanding contributions to India's strategic capabilities. It is none other than Dr. Chandrika Kaushik, Distinguished Scientist and Director General, Production Coordination and Services Interaction, DRDO. Let's welcome Chandrika Kaushik, ma'am, please. Thank you.

Lieutenant General Rana, CNC, SFC, Mr. Alkesh Kumar Sharma, former Secretary, MeitY, Mr. Subhash Gopinath, ladies and gentlemen. First and foremost, I must compliment the Bell and Global team for organizing this very, very relevant topic, at least from the security agencies' perspective, on AI for strategic defense, as part of the AI Impact Summit 2026, which is absolutely imperative today. Looking at the way things are changing, operational environments for soldiers are progressively becoming complex with the incorporation of advanced technologies into military operations. Be it network-centric warfare, multi-domain operations, unmanned combat systems, these complexities are resulting in increased information overloads, reduced reaction times, and quick decision cycles for the human operator. In order to cater to these challenges, it is the need of the hour to introduce autonomy into our military systems to augment the capabilities of the soldier.

In June 2018, the Government of India published the National Strategy for AI, focusing on how India can leverage the transformative AI technologies to ensure inclusive growth in all sectors of society, from sensing to comprehension to action. Today, AI has pervaded numerous end-user applications, ranging from computer vision to audio processing, natural language processing, knowledge representation, machine learning, expert systems, etc. AI is being classified into three variants: descriptive AI, which describes what has happened; predictive, which is anticipating what is going to happen; and prescriptive AI, which recommends what to do. With the emerging technological capabilities, AI is geared up to serve as a promising tool to provide enhanced situational awareness to the commanders and operators in highly contested battlefields, making our military systems more accurate and our military operations more swift and precise.

If we look at the various domains where AI will get integrated, and some of the areas the ecosystem is working on are underwater surveillance, navigation controls, advanced material characterization, hybrid high-fidelity models in the naval systems domain, trajectory and manual planning, target identification, threat classification, aerodynamics, and controls. So, I have listed out. I think for paucity of time, I will not be able to get into the details. But if you look at it, you look at armaments, combat engineering, missiles, strategic systems, electronics and communication systems, computational cyber systems, everywhere there is an application where AI is getting seamlessly integrated.

CICS operations with man-machine teaming in urban environments can be carried out using coordinated combat operations in complex, high-density environments where challenges of limited visibility, civilian presence, multi-story buildings, and ambush risks. All these can be handled by using AI-enabled advanced surveillance and integrated systems for enhancing the soldier's effectiveness, leading to very, very effective man-machine teaming. With increasing physical, psychological, and cognitive loads in highly contested environments, the future soldier and his equipment will need to be integrated with all nearby sensors and equipment. In such a scenario, AI will help in integrating the sensory information along with the commander's decision-making process through shared mental models.

Get Our Mobile App

India AI Impact Summit 2026 LIVE: Lt Gen DS Rana, DG DRDO Chandrika KaushikSpeak At Special Session

Transcription