Transcription
[Applause] Who was your best professor? Not your favorite, your best, the person you learned something from that's still with you today. Maybe it was something about history that helped you understand the past or a science experiment that helped you understand how the world worked. And then imagine this morning you discovered that was wrong. This is the quiet risk that is happening across education today. And the reason why is that traditionally we had three tenants of education. The textbook, the teacher, and the libraries. Textbooks were riddled with facts and information and academic review to ensure their accuracy. The teachers have dedicated their lives and their education to the very discipline on what they were teaching us. And then libraries were full of books and evidence and artifacts of our past. These were the core components to education.
But the reality is students and learners today have shifted from those core tenants and over to technology in generative AI. And that's where a core part of their learning is happening today. And one of the challenges with generative AI is that there is no bibliography and there is no accountability. But it is confident. It is instant. It is different in the sense that you don't know what's accurate and what's not. And that's part of the challenge that I think we need to address. What does learning look like when knowledge is actually the question?
I spent 25 years in technology and nearly a decade now at the intersection of education and technology and in that time I have witnessed firsthand what's happened in the transformation in the education industry. I've watched the shift from the traditional printed textbook over to digital-based learning solutions. Traditional textbooks are built on accuracy and they have peer reviews, fact reviews. Even when they're being purchased, they get reviewed yet again. And when there is something discovered that may be wrong with a textbook, there's a process to get it corrected. And today with generative AI that doesn't exist. So you have access without accountability. And that's one of the core challenges that I think we need to address. So the big question becomes how do we ensure that there's accuracy beneath generative AI as it's used by students.
I spoke at an event last year to 300 graduate students and their professors and I asked the 300 students how many of them had used generative AI in the last semester of their learning and they all raised their hands. I asked the professors how many of them had used generative AI in their syllabus and helping to teach these students. No one raised their hands. Felt bad because I felt like I had put the teachers on the spot in front of their students. So I talked to them afterwards and they said Jim, of course we've used generative AI, things like OpenAI, ChatGPT, they they've used these products but like the difference is Jim, I can't just use something and then bring it into the institution, bring it into the classroom, there's a process we have to go through. So we find ourselves flat-footed knowing that the students are advancing very quickly with this technology and we indeed are not.
So the reality is students are racing ahead. And one might ask how did that happen? The reality is it happened because the technology was thrust upon us and the early adopters were students and learners. And those that created some of this technology did not think about their primary validating users would be learners and educators. So we validated that these tools are fabulous for society and we didn't include the institutions who are now felt left flat-footed. So it's a bit reminiscent of the gold rush. And when you think about the gold rush, it had amazing risk and amazing potential reward. And up till now, you may think that I'm not a huge fan of generative AI, and it couldn't be further from the truth. Matter of fact, I think this is the technology and innovation of a lifetime. And I think the potential of what you could do with this is profound, not just in education, but in all kinds of applications. And when you think about the potential reward in education, the idea of personalized learning, the idea of you being able to spend some time with a topic anytime, day or night, and being able to go into your native language, if English is your second language, you could be able to do that with this technology today. And then we know, research has shown us that if all of us had a one-on-one tutor, we could see a full grade improvement as a result of that. But it's never been scalable, never been affordable. With this technology, that is possible. The notion of having someone that's available to you all the time that could help you move ahead in your education. And then finally, the scale and global reach. More than a billion people have used these generative AI solutions, meaning it's available to everyone everywhere.
But just like with the gold rush, there's a catch. In the race to introduce generative AI, including education, there was not a lot of forethought in how this would be delivered or introduced. It was simply there. All of a sudden, it was available to everyone. And so, we've seen this happen before with a revolutionary tool that was supposed to help us connect with each other and share information. And of course, that was social media. But the difference between social media and this is that social media is riddled with misinformation. It's everyone's opinions and ideas regardless of its fact base. And if you take that same principle and you apply it to this, then we are on a very slippery slope about where we could end up with this technology. And we know from those platforms that misinformation and conspiracies are far more interesting than facts. People share that immediately because they see the excitement and the enthusiasm about something that may be real versus that is real. But for generative AI, it is that this may not be optional like social media is. Once it's in the fabric of education, we won't be able to get it out. And GenAI doesn't know the truth. It knows the next best word. That's what it's trained to do is to understand the topic and the idea and creating what is the next best word that should follow. And experts predict that large language models will run out of data by 2026, just a year from now. And the reality is they won't run out of data. They'll run out of free data. And then you start to wonder what happens as generative AI produces data and results. And what happens as you start to feed other generative AI as its data source. So take a small problem and now it's magnified and it's in the fabric of this technology. You can't get it out after that. This is a real problem that we need to be concerned about. And what happens when generative AI follows the path of social media where engagement is more important than accuracy, where you prioritize clicks over credibility. It could shape learning even though it's not true.
So let's imagine a world just five years from now in 2030 and we've got two scenarios. The first is the free generative AI world funded by advertisers, optimized for engagement over accuracy, and the schools, particularly those in need, leverage the free solutions for all of their students. But we know that the solutions are full of unverified information and biased. The very engagement of the student, student and their data becomes the payment that fuels the corporate profits. Scenario one. Scenario two is a curated generative AI world. One that is fact-based, that is as accurate as a textbook but available to everyone and it becomes a public good that every student has access. Both options are viable. You have to decide. We have to decide which one do we want.
I want to share another story with you. I presented um at a meeting in Georgia with deans, provos, presidents, and tenured professors about generative AI. And one woman took the liberty of changing her name tag and writing the original AI equals academic integrity. And as I finished my talk, her and I spoke and she was very clear with me that my enthusiasm and excitement about generative AI for education was going to destroy academic integrity. And at the time I was really dismissive to her. In hindsight, I don't think I believed her. And I think that my bias, whether it was my enthusiasm about generative AI and its potential or whether it was my personal technology bias, but the reality was over time her concerns were valid. She may be more right than I was. So if we do nothing, students won't know the difference because the tools behind them won't know the difference. So how do we shape AI and generative AI into a tool for both learning and knowledge?
So let's talk about a way forward because there's nothing worse than a problem without a way forward. So let's start with agentic AI. It is all the rage. It is available. It is working today and it is reasonably easy to configure and set up. The notion of having expert agentic AI on specific disciplines, specific topics, it is something that can be done today and is being done for other things, not just in education, but it is a viable option for us. And then let's talk about cross-sector collaboration. The idea of getting the people that care and are concerned are involved in this ecosystem together for a conversation about what can we do about the very problems we're talking about with the goal of having fact-based learning, not just on engagement and metrics. And if you think about it, we need to have some certification around generative AI in education. We've seen this happen with the FDA. We've seen this with the Privacy Act in GDPR that protects our information across the internet. These things exist and in many cases everyone said that they couldn't, yet they do. I think the opportunity is here for us to do that so that we could have an agnostic body that validates what is a safe, viable, and accurate generative AI solution for educators. And then finally, making sure that all of these things are peer-reviewed just like academic research. Again, this is possible. But the reality is we have to begin now. This is not something we can wait another decade to do something about. We need to begin right here, right now.
Generative AI is it's not a teacher. It's not a textbook. It's a mirror. And whatever you put in front of it is what it reflects back. And so far what we've put in front of it hasn't been great for education. And I think that's a problem that is worth solving and a problem that is solvable because if we build that the way it is today, the very foundation of education I fear could be inaccurate for a very long time. And I think students today should trust the answers and the knowledge that they're gaining the same way that many of us did that had those core tenants of education I shared in the beginning. The teacher, the textbook, and the library. But if we don't act, generative AI in its different forms will just continue moving forward without us. So the future of education is being written right now. The only question that remains is who is holding the pen. I think it should be us and I think it should be fact-based. Thank you.