Transcription
Well, welcome to our panelists for this next segment. This is on AI's report card. You know, in ten days, it's going to be ChatGPT's third birthday. So who would have known three years ago that this little textbox would turn into an AI boom that would lead companies to spend trillions of dollars? But here we are. And it really shows we don't know what's coming. When things happen, they can take us by surprise. So I'd love to to speak to our panel now about how it's all gone so far. And I'd love to start with Mohamed. He's the chairman of of PricewaterhouseCoopers. We know the AI adoption has been so fast for consumers. ChatGPT is used by 800 million people a week, but it's been slow for businesses. Now, you're part of your work as helping businesses implement it, and you're also using AI a lot yourself at PwC? So what are the wins and what are what are the struggles for you guys?
Very interesting question. I would tell you that when it comes to the wins, I think we are past the stage where companies, including ourselves, are asking whether they should adopt artificial intelligence or not. That is no longer the question. Right. The question is now, how fast can we go? So some of the wins that we have seen where, you know, companies using AI to become more productive or more efficient, we did it to ourself with the different people, the people that we have across the world. And this has been a success. The first thing that we had to do is to make sure that people didn't fear artificial intelligence for the jobs, because it's one thing to use artificial intelligence in your personal life. On the job. There were some threats around job replacement for artificial intelligence. Yeah. So some of the wins that we have noticed are companies using artificial intelligence and making sure that they have the people embracing artificial intelligence. I mean, one of the first decision that I had to make was to decide how do I make our people learn about artificial intelligence on the job? We give access to all of them. In all 139 countries. And I had to ask my question. What do I do in the U.S.? What do I do in Europe, in Asia? And that was the one question. I said, give it to all of them. To everybody. Just give it to everybody and let them play with it. And we saw a lot of the productivity improvement coming directly from our employees. And as you and I discussed, we even told them, do not think that artificial intelligence is the threat to your job. Consider artificial intelligence. As a digital colleague, it flips the mindset of the people for adoption. Yeah. So there are some of the window to have seen some of the struggles was around How do you measure the return on investment? Oh, yes, because you see the adoption on the productivity side. But in many companies, including us, this has not led to letting people go. Attrition or replacement, it was more augmentation. That's what we have seen. And augmentation sometimes is hard to measure because it's the thing that you choose not to do rather than think that you stop doing right. Yeah. So that's that's what we saw. And now we are seeing to it before artificial intelligence. Things are moving into how do companies use artificial intelligence to change what they do to do things differently or even do things that they never did before. So that's where the struggles are today. But I would tell you that we're in a very different place. You save almost three years for ChatGPT. Mm hmm. We are in a very different place about asking our question of whether we should use it or not. Now. Yeah. How fast do we go? It's a no brainer. It's a no brainer Now, by the way, it's not that the future is coming. It is already here. Yeah, well, we're. You're building it. Yeah, exactly.
So, Forrest, you run and have founded Sea Ltd, which essentially the Amazon of Southeast Asia. That's how some people refer to it as the e-commerce giant among other software and online services that you provide. And Asia has seen a huge adoption of AI. How are, how is your company using it and what's worked and what hasn't worked so well?
Well, I mean, as a technology company, we're super, super excited about what we have first thing in terms of the AI advancement and the development and like we would admire what has being developed in Silicon Valley and the rest of the part of the world in terms of the fundamental technology breakthrough. But as a technology company is basically working like a like a focus on the individuals with consumers, and especially for our emerging markets, specifically in Southeast Asia and Latin America, we are more excited about what that means to people's daily life. So and we have seen a lot of impact right under and I would describe like a the AI is a kind of a very, very powerful production tool at this moment. And I think like for us, we are all grown up. We are familiar with the PC, with the laptop and the probably more recently that the smartphones, just the imagining how much less you can do on a daily basis without access of a PC and without access of your phone. And I mean, a thing like that is a way I think about like a definitely the new tool, the new powerful tool as well, just to make everybody's life daily life more productive and and probably more enjoyable. So that's the like because of the PC, because of the laptop, my life is a much, much more productive and and enjoyable, supposedly in e-commerce. I think I love to give audience a bit sense about what we have. Thing is a really, really working and we are serving. We are not only serving hundreds of millions of consumers because we are also serving like a tens of millions of sellers, right? They are. The order could be individual, could equip, could be those small medium sized businesses. Right? There are entrepreneurs and they are selling things through our e-commerce platform. SHOPEE And are the things that a small example could be like. Even I'm a good seller, I can get my merchandise in like a more cost affordable, right? And but I do not know how to introduce I'm never really educated how to write a product description. Yeah. And how to take a nice photo because that is a require very expensive skills and how to shoot have a nice video and up. I may need to hire agency and hire somebody to do the work and the quality may not be always there. And also depends on how much I can afford. It depends on my what I sell, what is my margin. And now like a we just from the platform perspective, we just provide those tools to millions of like a like a sellers right on that they can they can easily just ask us to use our tools and to put it, but just a picture, whatever you have, and that the tools will kind of generate a much just that will touch off, just a mimic, become a nice much, much nicer photo. And what I even can do is that it just a come by several photos into a quick video and it's a very, very useful for when buyers look at like a now I don't know I never read like a text of menu about how to use a product like, okay, I always just go to the YouTube search, okay, how to use these things like just will watch a one minute video and I got it right. And I think that is a kind of the similar experience on e-commerce platform and people kind of doesn't want to read a product like description or like just to look at photos, a quick video that helps.
And for us, can I just ask, since you've implemented that, have you noticed any I know it's early days, but have you noticed any difference in sort of your revenues or the number of products being put up on the site has as that feature actually help the business in some way?
Yes, we provide this kind of like a for free to all the sellers because with. Is that like a you generally like a benefit to everybody, both the sellers and the buyers of buyers. I also love to see nice, nice photos, write a nice, nice content and clearly we see the adoption is very, very high, which means it has increased as the seller as a seller because they can choose to use or discounts was not due. Right. And the take rate is very, very high. And on the buyer side and we realize the engagement level is going up and the conversion rate is it's going up like they stayed longer time with the product. They want to know more and these quite enjoyable experience of watch a video and the and and also like there's more conversion rate on in terms of the placing order and another part from the buyer perspective is like a you already people love to read reviews there could be for one product, one seller people care how other people think about that could be hundreds of reviews and a and is to take time to read all of the major to process okay which make more sense which is a more relevant one do I trust. Yes. And then now there is a kind of like a quick AI summary. Okay. So this is in general, how many people think about this is a good seller, this is a good product. And now if you really want to dig into details, you can still read a review Better as summary can give you a quick kind of a sense about about the sellers and the product.
That's such a great kind of lay down a rundown of how that's working for you because I think one thing on the minds of so many businesses is return on investment for AI, because it is expensive. Of course it's expensive for the companies to develop. And as you Mohamed mentioned, you can see that there's productivity, but that's perhaps a little bit difficult to measure. In Forrest's Case, that seems very clear. You've got very clear metrics that you can measure government. How on earth do you even measure the return on investment on I guess you're listening to forests. Absolutely brilliant example of how he's using it is a good reminder to me. And people in government, you know, our primary job actually is is just to create an environment where people like Forrest can go and innovate and do what they do. And so that's when we are conversation earlier about regulation. Well, I've always been in the camp of saying that we must be pro-innovation when it comes to this new technology. We're in the very early days of it. What we do not want to do is, is do anything that will stifle Forrests and entrepreneurs like his ability to try new things to generate value for for their for their customers. So I think that was a great reminder on that. On government, the most the most easy thing to do is look, how can we automate all the government services that every one of you have to interact with on a weekly basis, daily basis to just make that both cheaper and more efficient, which particularly for constrained European economies, is is of paramount importance right now to find some savings and government spending. But how can we just make life people's lives easier and for us also mention that from a commercial perspective, But government should set itself the same bar is, you know, how do we just make those interactions with government easier and make people's lives better? And give you one practical example, One of the initiatives I'm particularly passionate about is helping introduce a targeted screening program for prostate cancer in the UK. Now, prostate cancer is become the most common cancer of men in the UK. A one in eight men will be diagnosed with it. And I've been advocating for a targeted screening program, national screening program, because prostate cancer, if you identify it early, we're very good at treating you and keeping you alive. And the problem is it's symptomless. So it's if you miss it and you only pick it up later, then your chances of survival drop quite precipitously. Now, you know, obviously, introducing a targeted screening program, countrywide basis is quite an expensive and costly undertaking. However, I it has changed the dynamics on that because now I can read all those MRI scans that are part of the diagnostic pathway. I'm sorry, by the way, you mean generative that. Yeah, the latest tool, the latest classical machine learning stuff. This is the real opportunity which is which is meant that whereas before you think, well look, do we have enough radiologists, can the system handle all of this Now it's very clear. Yes, it can, because these new tools are incredibly accurate and quick at assisting our clinicians going through the MRI scans, which now makes it much more easy to say that we should introduce this targeted screening program. We can afford it. We can handle it from a capacity perspective and in the process, then save lots of people's lives and and keep people's loved ones alive. Now, you know, those are kind of the practical examples where people can say, okay, this is making a difference to me. And this is, you know, the government, the state has helped make that a reality. So I think trying to find those examples is is really, really important beyond just saving money so the return on investment can be saving at. For human life, for governments.
So the title of this panel was the AI Report Card. So I'm going to ask you all a quickfire question, and I'll start with you, Mohammed. Since you were one of the first power users of generative AI. PwC how what kind of grade would you give generative AI in terms of how useful it is today with a being the best and F being, It's a total flop.
I give it an A.
Okay. And why is that?
And I would tell you why I give it an A. It's not because of what we see today, but what might come. We get a truly believe that artificial intelligence. He's going to create a new age of innovation. There's so much capital. Being invested. The example that you just gave, you are innovating. How you running your business. You're talking about, you know, vetting around screening for diseases, etc.. What's coming next. So when I look at the potential today. Of artificial intelligence, not just generative intelligence. You have a, you know, industrial A.I., artificial intelligence coming in and all of the things that are currently in development and the amount of capital that being deployed. We are only witnessing the beginning of a new innovation age. I truly believe that. And the last time we saw that was 25 years ago, and the amount of capital going in is a reason you're optimistic. The fact that so much and the speed of development that we are seeing, the way the models are improving really, really fast, the looking at all the use cases or the agents that are being developed, you know be Rishi and and for us get from exemplified to being used every day we have seen new examples not based on what is being provided the large language models. Yes I know a lot of innovation. You know, the new ships that are even more powerful than before. But I would tell you, I totally agree with you. The supply side of A.I. providing A.I. is a great thing. We need it. The adoption side of AI is what's going to create the future. The diffusion that you talk about. Diffusion. The diffusion is important because that's when you create new models, new business models, new companies, new ways of solving problems. And that is why give it an A, not because doing what we do today more efficiently, but about the things that we would be able to do that we cannot do today.
Okay. So Muhammad Forrest Muhammad's great is based on the future. You can do that, too. Or you could give it a great today. So please, what would you say?
Well, I would say I will give seven and under in the context is I'm I'm truly excited and very convinced and AI will be a great development and will create a lot of lot of values. And it will change people's daily life in many, many aspects and in general, I think make people's life better. And but I think that the reason I give a little bit lower score on Mohammad is I feel we are still at very, very early stage. Yeah, Yeah. So we shouldn't be at all satisfied what we have thing and compared to what we are going to see in the next three years, five years or ten years, I think like a ten years later when we truly appreciate and, and, and what I can do and can change the society and to communicate, change in life, whatever we have done today is very, very small. It's kind of like it's it's yeah so that's a that's it's a still I think is the reality well beyond today what we can look what what we can imagine so that is a lot is a my my my take and, and specifically I think like there will be many, many kind of like a aspect, right. And there will be change. And the, but today in our organization, we just completely embrace the AI. So we like from our employees state they ask us what is our company is the kind of the view and a standpoint I should come we use all the all the tools and what is the consideration in terms of the safety and that conflict, confidential concern, all those things. And in general, we our approach is okay, we just embrace it and that you do and we try to figure out all those things by working closely with all the big big, like a giant, like a Google or like a openAI right to it, to embrace you to adopt their their models. And at the same time we figure out all those the issues bearing people's mind. But in general we say, okay, do it. And if it is something you don't work out and we will, we'll fix it. And at the same time, in in, in company, we also developed our own kind of like the agent platform. So we provide a tool that the productivity tools for all the employees and even our operational team without any coding experience. And if you can develop their own agent and AI to do this today or like a day to day work, solve their pinpoint, I can give a quick example I think everybody can appreciate right under we handles of like a million stuff a puzzle every day and the e-commerce business and we always get concerns like the people say oh I didn't receive my puzzle or like a and a how come I yeah, you supposed to be here But I didn't. Every time we get those a customer complaint, we have to investigate it. I know. Is it you did for the wrong places or like a or the maybe the delivery driver steal it or like a to the customer just to kind of give us a force report. There was a so much, much work. And if it's the individual case you'll make several phone calls is handled. But if you think about is a million of like a yeah. On that skill and the total skill and even just a 1% 2%. So that is a lot of a lot of work And our Malaysia local team logistics team and the using our in-house agent platform to develop an agent. So basically every time like when the deliverymen deliver a parcel to the customer's home that they didn't, we need to take a photo as a as a proof and, and then upload before they also upload to the to the to the target. Database, but now they upload and then the agent can immediately give the feedback to the photo is not clear enough or like it is is a is a is another as a good profile, you should make the address. And these are clearly show this is a parliament or this is just a a house right. And the, and and the, and give the. So this is a resolve a lot of pre resolve a lot of a dispute and people say okay I submit the photo as approved but the customer say I didn't receive it and and it's like a whole other workforce that you've got doing that for you. Yes and it's a pretty much it's it's a it's a pretty much a real time. So that's a that's a solve. A lot of a lot of a lot of a lot of a problem. And I think these are we have a many, many tools again so this is that the tool empower people without any coding experience and it has developed a such a software tool. Yeah. And like relatively quickly and to put into daily use is a solve the practical issues they have today.
Ask a nice example Rishi what's your what's your view on I, I yeah I tend to agree with with with Forrest but you know listening to him you think isn't that an extraordinary use case already for a technology that as you said is is barely a couple of years old. And I think that's the exciting thing. That's why Mohamed is giving it an A-grade today is because the pace that it is improving at is really quite striking. Yeah, and that was the thing that I found in the job. And it's tricky for policymakers or any of us to kind of understand exponential growth in general, but the speed at which these models are improving themselves and then, you know, you kind of think through what that will mean in a few years time. What grade would you give it?
I probably where Forrest is. I see. Yeah, exactly. Exactly. What do you you're you're the optimal. But I would say that I totally agree with the two comments but, but we are very early. Right. This is the point. We are in the foothills of this and it is improving so quickly. Yeah, that can you can feel very optimistic about where we're headed.
Well, Mohamed made this point a very important point about capital and how much is being spent. And for you, it's a sign of reason to be optimistic. The fact that so much is being spent, particularly because in the past, when a lot was spent, then it led to good things. But this is on present precedented, trillions of dollars being earmarked for data centers and chips and other forms of infrastructure. And what I think is really paradoxical for a lot of us and probably for you as well, is it's not really within our control. We will be affected by it. But these decisions and this money is largely being spent by a handful of very, very large companies. And some would look at that and say that's a real consolidation of power happening economically and terms of influence and of impact if something goes wrong or if someone with a crazy radical agenda gets becomes in control, that could we might pay a price for that. And I just wonder how problematic is that, that concentration of power that we're seeing in AI, or is that status quo? Actually, it's worked out fine and it should continue. And I put that to the panel.
You know, this one, this is a great question because we think about, well, there's only a few companies that have the capital to do this. Right. Cost a lot of money. And in artificial intelligence, we all know it's not just the development of the models build out of the data centers that you mentioned, buying chips, cooling equipment, industrial equipment, real estate. It costs a lot of money. You don't have enough companies in the world that have the balance sheet to do that or the capital to do it. And by the way, you know, you cited a couple of the thing that the printing press now do, the industrial revolution and all this. It would not that many companies had built it because too much capital. So when you look at the signals of today, they are the same that we saw before. We couldn't just intelligence age with artificial intelligence, but the same fact pattern happened before. So that concentration is okay on the supply side. Okay, now on the demand side, on the diffusion side or adoption side, that's where you will have creation of new business models. New companies are going to come in. People will change how they do drug development, how for it is running its business today, you might actually build some new business lines. We could offer artificial intelligence, too. It's okay to have concentration of power on the supply side because it would create innovation on the demand side. That is why I don't worry about this as I'm going to keep spending and you point to run spending. You know, a lot of questions were being asked to us around. How do you. What's going to happen next? How do we measure return on investment and be doing all the things we were trained for the past 15 years to have the answers? You know, if you were doing the following things, some of what will happen next. Yeah, the world is changing. It's uncertain. Nobody can predict the future. So what do you do then? You look at leading indicators because you don't have the lagging indicators, and the leading indicators are speed of innovation. Yeah. Capital being invested. Yeah. A number of use cases that people are trying. And if all of these things happening, something good is coming on the other end, including innovation for companies like Forrest is running today. That is why I'm so optimistic. But I would tell you, I don't fear that concentration. I'm actually glad that they're spending the money because we are all going to benefit from it on that point. And feel free to talk on this. The other question, but people have talked about an AI bubble. It sounds from all of you like that's not a huge concern because this is all being steered in the right direction or am I wrong in saying that?
Forrest, you go first.
Well, I mean, I think it's a it's like a how do you define a bubble right on the end? I think it's like a considering what value potentially could be created if things are totally fine because it's like, as you mentioned, and now the key drivers of of this, a new technology revolution, is a by several big giant companies in the word technology, giant companies, and that there are all the commercial companies and the and they all calculate the return right And they are all very savvy and very disciplined in terms of every single dollar to invest because they are like a I mean, I have a lot of interaction with them that they have been all in the public, like most of them, the public company, for many, many years. Right. I never built up this a discipline and and a capability and how to deploy capital, how to do the capital allocation efficiently. I think if if they believe okay, so as you mentioned, is like a hundreds of billions of dollars have put into this new technology development on an annual basis and that the returns that is like huge. And I think like from the talk about a bubble perspective that is a is about like what time horizon you look at and you think it now is like a things is a short of a datacenter a lot of investment go to the data center, go to the hardware that is a go to the electricity, go to either the power to water. It may be like in the short payoff time, easy to perfectly pace with the technology development and the end of transfer to the real value of a people's daily life. Maybe there is a had then that there is a behind right hand, but in the short term there is always is an adjustment, a line it to happen. But in the longer term. Right, I think those existing structure we built today will eventually benefit and the whole society and the technology never disappoint human being in history. Right. And we always believe technology advancement is the single biggest value creator in human history. Right. And from that perspective, I think I think there is a no bubble. And and but that doesn't bubble from that from a long term perspective, but doesn't mean. Okay, so you buy AI related stock and tomorrow you're making money so that I'm not commenting from the capital market perspective but that from the technology that eventually the value creation, I think that's fine.
I think there's an interesting thing that Forrest said there, which is about, you know, historically we've always benefited from new technologies and that has worked out and created a lot of positive impact for society. The the interesting thing, if you look at it today, is amongst Western countries, and she says that I Anglosphere in particular, there is actually an enormous amount of fear and skepticism about A.I.. There's a lot of fear and that is interesting and it's different to what you're seeing in countries like China or India or the Gulf or here in Asia more generally, you know, where people have just a different view about it. You know, they're more positive about it impacting their lives. They feel good about using it. You know, you are not seeing that in in the West in the same way. And secondly, it's different to the past because if you look at the last digital revolution, you didn't see that. And so when when we had the previous digital revolutions, people in even in the US and the UK were very positive about the technology and they are not in the same way today. They are more fearful. And I think that is a challenge. And also companies who are driving this fear missing out. So that's just another type of fear in a way that's also drive growth and that's where you're getting the investment from. So I think in order, you know, in order to make sure we get all the benefits that, you know, we're all talking about, I think it's incumbent on actually everybody. It comes from political leadership, but also from industry to be able to bring their populations in these countries with them. Yeah, because actually that that to me is the thing that I look at and I think it's all great and we can sit and talk about all these wonderful things, but you've got huge swathes of your population that are skeptical or nervous. We need to overcome that. Yeah. And I think that that is something that, you know, I think a little bit about and a jobs are obviously a part of that. But it goes it goes more broadly than that. And this is happening very quickly. And I think you touched on this at the beginning and you look back at these technological revolutions. It took you know, when you look at how long it took 100 million people to to use the technology, I think it took something like 50 odd years for 100 million people to to use electricity. Well, to have electricity. And you fast forward to the PC. I think it took something like 15 years. Yeah. And then the Internet with seven years. Right. Chat GPT was two months. Right. Like that is the speed that this is happening around the world is quite unprecedented, which is exciting. But it does mean that. We have to work harder and quicker to bring people along with us.
One final question. Does anyone want to just answer something that address what Rishi said about the fear that people have? And how would you address overcoming that fear that people have of AI? Is it about just showing them how it works or is it a policy move or is it a cultural thing?
I would say transparency, governance and also a set of regulations that are being developed that people that actually understand artificial intelligence. Yeah. Yeah. Great answer. Okay. So very optimistic panel here on AI, but also with some concrete points about how things need to improve. So thank you so much to our panel and thank you very much to our audience for listening. Thank you very much. Thank you. Thank you.