 Mina Alaray, the editor-in-chief of The National, based here in the UAE, and it's my pleasure to be moderating this plenary session with you today. So the question being asked, is AI a force for good? Well, it depends. And we have the next 30 minutes to look at the elements that will either make it a force for good or one which we should be worried about. It has touched the life of everybody one way or the other in advanced economies, and of course, more is to come. From chat GPT to the future of work, large language models, supercomputers, and the intelligence of machines will touch our lives and usher in a new era increasingly. And yet threats exist. Those threats are not only impact on society and work, but also cyber attacks that we see increasing. So we have Azim Azhar, who will do a much better job than I, to tell us about AI and how it can be a force for good. Azim is the CEO of Exponential, based in the UK, and then we will be in conversation with Azim and his Excellency Khalfan Belhol, CEO of the Dubai Future Foundation, and we will want you to be part of that conversation and present us with your questions. So in the meantime, it is my pleasure to introduce Azim, please. Salam alaikum and good morning. Your Excellency's distinguished guests. This is a very big question when our host at the World Economic Forum asked me, is AI a force for good? I thought this is a very, very large question to tackle. We've heard many warnings about artificial intelligence and Mina mentioned some of them, the risks of disinformation, the risks of lethal weapons. But I would like to be clear in my short discussion today that even when we talk about artificial intelligence and just through the lens of its benefits, this is a very, very tricky question, whether AI is a force for good or not. But to get there, let's start with an even broader question, which might illuminate where we go in my discussion, which is, has technology generally been a force for good? And to that question, what is technology? What do we mean by it? At its core, technology is the compounding of human knowledge, generation after generation, each technology building on the knowledge of the previous years. So the tree of knowledge that we see is a tree where the more advanced technologies sit on the layers above us. Ten millennia ago, our ancestors started down the process of cultivation. And that cultivation turned into agriculture and farming and intensive farming. We started to be able to feed populations. And as a result, our species has done reasonably well. The Babylonians a few thousand years ago developed chain pumps and sophisticated irrigation systems. They managed to transform their arid environment into a sophisticated and complex society, featuring societal growth and political stability. Now, thousands of years later and countless technologies later, the steam engine in the 18th century in England was designed initially to address the practical issue of floods in coal mines. But from the steam engine came the railway system and mechanization. And with that, the industrial revolution, which really gave power to England and it brought us and heralded the modern world that many of us that we have all experienced over the last few hundred years. So in some sense, technology is essential. Perhaps it's the essence of what it is for humans to be humans. But now let's think about artificial intelligence because, of course, artificial intelligence does not exist as a set of technologies independent of other technologies. It's not here in isolation. It builds on the millennia of work that went before it, whether it's advances in mathematics or computing or chip manufacturing. It's dependent on that layer of technology. So in some sense, when we ask the question, is AI a force for good? We're also asking the question, is technology in general a force for good? But there is something that is quite distinct, I believe, in artificial intelligence that makes it a different beast, a different fish to the technologies that came out before it. And that is that the AI systems that we see today compact and combine all of the previous knowledge of humanity in their models. If you look at something like GPT-4, which is a very, very early AI system, it's trained on three trillion words of text, which is if you have teenage kids, you probably see them on Instagram messaging the whole time. That's more than what a teenage kid can produce even in a week. These large language models encapsulate all of our previous knowledge, and they make that wealth of information accessible in mere seconds, millennia of information in the palm of my hand. So AI is not merely another technology. It's also a meta technology that encompasses and utilizes the full spectrum of our knowledge to the extent that it just sits in this device here. And it's a reflection of all of that advancement, and it allows us to query and generate that knowledge in new and rapid ways. That speed is a really critical aspect. The speed is distinctive. When the American president Abraham Lincoln was assassinated, it took 13 days for the news to reach the prime minister in London. And today, of course, those delays have been eliminated, and as AI systems become more sophisticated, they will be able to process that type of news much, much more rapidly in that millisecond scale. And we're seeing with these very early technologies, these early AI technologies, just how much impact they can have. In the past month or two, there's been a survey, a study from Harvard University, Harvard Business School, and the American consulting firm BCG. And what they did was they took several hundred of their consultants, and they gave some of them access to chat GPT, and others did not get access to chat GPT. Now, you have to understand that BCG consultant is amongst the most sought after employee in the world. They have to graduate the best schools right at the top of their class. And what they discovered in this study was that the bottom 50% of BCG consultants, who by any measure are excellent employees, improved their performance on strategy consulting tasks by about 40%, which is remarkable, using a free piece of software, chat GPT. The top 50% incidentally also improved their performance, but not as much. So I know that within the members of the Global Future Councils, you're all in the top 10% of performance. And just the bad news for you is that chat GPT is going to give you some new competition. This technology, of course, can also catapult our decision making, our ability to do fundamental science. Now, while it's not a large language model, DeepMind's AlphaFold is a technology that allows us to explore the structure of proteins. And in just a short couple of years, AlphaFold has been able to map out the structure of all of the millions of proteins that science knows about. A scientist using AlphaFold works about a million times faster than a scientist who is not using AlphaFold. And the speed with which this information processing can happen, we see in more and more domains. There's a particular type of research called a meta-analysis, which is common in medical and healthcare research. It's extremely powerful because medical research, when you have to test things on patients, is slow, it's time consuming, it's expensive. And so what meta-studies allow you to do is they allow you to look at all of the different studies that have been done on a particular condition with a particular drug and aggregate the results and deliver new insights from the existing body of knowledge. Trouble is, meta-studies are time consuming and expensive, so we don't get to benefit from them. Now, there's a new AI tool called Illicit, which is built on a large language model which dramatically shrinks the time it takes to undertake a meta-study. And so what you see is this time compression and these silicon-based systems fundamentally outpace our ability to work. At any point that we as humans are slow at making decisions, slow at conducting the structural analysis of a protein, slow at gathering the information for a meta-analysis, slow at analyzing the market prospects for a company wanting to launch a credit card in the GCC, an LLM or another AI tool can speed up that process. And so these tools end up being very, very general purpose and finding their way right across our economies. And that's why I say they're disruptive by nature because they're disruptive as sure as if a soccer team hires a brand new talent that hasn't been seen before and starts to win match after match after match. And I'm sure Hurt Harland is a great example of that with Manchester City. So when I think about artificial intelligence and its context within our economies, I think it is ultimately the underpinning of the modern 21st century economy. And it's going to have two components. It's going to have an infrastructural component that is the layer upon which other services are going to be built. And it's going to have an interface element that is the lens through which we access all of our digital services, whether they are public governmental services or private services. Because they're so vital, it turns out that those are positions of power because if you control the infrastructure and you control the interface, you really control what can be delivered and how people can access it. And so AI ends up being a question of power. So when I look at the question, is AI a force for good? I think it's maybe not the right question. Maybe the right question is this. If control of infrastructure is power and control of the interface is power, those who control artificial intelligence will be powerful. So we should ask questions like, should the decisions about where power lies in a world with artificial intelligence be taken here or there? What are the architectures of power that foster greater or lesser shared prosperity? And what arrangements will make sure that the economy delivers on what it should do, which is deliver human needs? Now, there is history to guide us on this. In the field of the economy, what we know is that competition is what drives this. Free market entry, the absence of oligopolistic power, the reining into the scale of firms are tried and tested strategies. In the field of politics, what matters is the separation of power, the ability to keep a check on bureaucracy, on government, on regulatory capture, and ultimately the participation through whatever mechanism is politically and culturally appropriate of the people for whom this affects. But when we look at the technology, it's not just about a question of power. It's also about a question of the speed. And one of the questions that we have to tackle is how do we think about our culture and values in a world where the speed with which these systems are going to change is going to increase and continue to increase? Because many of our cultural norms and our values are based upon the world that we've come from, which is that things change but not as fast as they are about to. And I think one of the challenges we have to address is what are the languages that we start to use in order to stay comfortable in a world that has a sky that will be even busier than the skies above Dubai International Airport? Because when all is said and done, these are all questions before which we don't have answers. Open AI didn't realize that chat GPT was going to be as successful as it was. No one would have realized that there would have been a generative AI boom in 2023. Had you asked the question in July 2022? And yet despite all the drawbacks of these systems, the hallucinations, the errors, they're proving themselves to be very, very useful. And so that's why I'm asking lots of questions because we don't really have answers. But what I do feel comfortable about is that even in some picture of the future where we have a modern, beautiful city, I'm sure we're all familiar with one of these, with futuristic infrastructure, when the day comes to a close, what we'll want to do is gather with our friends and live at a pace and a speed of humanity and human connection all of us together. Thank you. Thank you very much for those thought-provoking remarks and this idea of going back millennia to the Babylonians and yet thinking about speed with which now everything is being compounded together. And this idea also that technology is not just AI, it's everything that humans have done and put together. So we'll talk about some of these particular topics. We'd like to take a seat and I'd like to introduce also his Excellency Khalfan Bilhul to please join us. Okay, so Khalfan, I'd like to ask you first to reflect on what we just heard as in present, but also think about what we heard about the benefits and the productivity benefits. There's also the risk. How do you balance the two, especially from your vantage point? Thank you very much. First of all, I'm honored to be here and honored to be with my fellow panelists. We had good to see you again, Mina and Azim. Your job, what you've done is amazing and looking forward to continuing our conversations that we had. Of course, I'm reflecting on your presentation which obviously covered many aspects on AI. I really appreciated the speed factor and what you've mentioned and I want to elaborate on this a bit more because even as someone working in the government and working in an agile government, I cannot agree more with the fact that no matter how fast you are as a government and how agile this delta between how the policies and how innovation is advancing will always be on top. It's just how much you actually mitigate and shrink that gap. That's how agile you are. Because to your point, whatever you have adapted to or tried to come up with policies today, come two weeks, three weeks from now, there's something else coming up. And the best example I can give is GDPR and the efforts that they are coming up when it comes to data. And suddenly a beast of an gen AI comes up to us and now everything needs to shift to adapt to something like this. And many of those rules that were in place which were relevant to maybe data protection laws in many ways and forms now need to be completely reconsidered. So this shows you that an entity like this really needs to adapt. So we're blessed here in this country to be honest because leaders are literally a phone call away. And policies are a phone call away and sometimes a WhatsApp away from a conversation. And I'm being realistic. Sometimes you have those conversations and you need to adapt policy. And I'm also blessed. I feel very lucky to be in an entity that really tests for future ideas and has a top board chaired by the Crown Prince and ministers in DG that help us change policies as fast as we can. And yet that gap will always be there. It's just a message from us and no matter how, I mean, you just need to shrink that gap and ensure that basically you are there in the right time to change your policies. And when it comes to, of course, gen AI, and I'll follow the lead of someone I work very closely with and is dear to me is Excellency Umar Al-Ulamma who is our champion in artificial intelligence. And he came up with, I mean, two weeks ago we had a gathering called the Genitive AI Assembly. It's one of the products of Dubai Future Foundation where we convene with people on the most important topic. And clearly Gen AI was the topic this year. Metaverse was last year. And now I'm all confused. You know why? Because I'm in a council for the Metaverse today. And I'm discussing Gen AI right now. And my mind is all over the place. They're completely two important topics. And the first thing we said today for those members that are with me in the council today, the first thing we said is today in the morning. So Metaverse is important. But what if it was a Gen AI enabled Metaverse, which obviously will be the case going forward. So we need to, it's a sign of connecting the dots between different efforts. And it's a sign of alignment and unity that we need to cross collaborate and unite and work together. So we're going to open the floor. If you have a question, please just raise your hand so we get a mic to you and introduce yourself, your name and affiliation. So I am on the lookout for anybody that raises their hand. But before we go to the audience, I want to ask you about the near term benefits that you see of AI. Because it's really hard as we're talking about speed and how we're making leapfrogs every 12 months or so. In the short to medium term, what do you see are the benefits of AI? As I'm going to start with you. Thank you. And just a brief comment on how funds are observational. I mean, I think it is an incredible strategic strength to be able to respond very quickly to these changes. The only other organizations that are able to respond that quickly are the big tech companies. And the big tech companies will deliver the early economic benefits of Gen AI. So the obvious place is in increasing productivity within white collar work, which is 50, 60, 70% of different economies, whether it's working in insurance claims underwriting or business planning or marketing, these tools are sufficiently good because you can improve productivity in areas where you can allow for the occasional mistake because there are other humans. Ultimately, I'm sorry if there are any marketing directors here. It doesn't matter if there's a mistake in a marketing plan, but it does matter if there's a mistake in a healthcare treatment plan. And so I think that that's where we will start to see really, really rapid uptake. And in particular, while I don't think it's showing up in top down statistics yet, when I talk to execs in companies, everything from small startups to firms with hundreds of thousands of employees, they already have these tools in use in these non-mission critical areas to improve productivity. Can I just jump in here? And I think, I mean, I absolutely agree with this comment and going back to our example in the United Arab Emirates, I'm maybe referring back to Zexensee Amar again, because he's a perfect example on how he became a leader in a space that we had no clue about. And now he's, and this shows you the DNA and the setup of the United Arab Emirates. We just jump into things early, and an individual that's hungry enough is identified to be the expert for this space. This is a space that we don't know, but it's a question of when do you jump in? And I remember him speaking in this session being, I think that there's a Spanish AI minister now, a woman, which is great also for diversity. So it's being picked up at the moment. But the idea here in the DNA of the UAE, something important from humanity, something that will add value, not for Dubai D, but for the whole world, let's convene and convene early. And that's what happened back in the then in the cabinet when Zexensee Amar was appointed. If you look at the blockchain, we had no clue in 2016, and we had a conversation what blockchain was. We convened, we had a discussion with understanding it a bit more. I won't say we're experts in it, but we had, we created setups and now you see the big companies setting up shop here. So the idea is really to create a platform for hope, create an environment that's inclusive. We have more than 200 nationalities living here, 90% of the population are the expat community. So we intermingle and understand the different concerns on a global basis. And then we jump on that space on it early. We can make mistakes along the way and that's fine as long as there are risk-mitigated mistakes, putting safety and health as a priority. You cannot launch, I know I'm going a bit off-topic me now, sorry, but it's the concept of embracing technologies. You cannot adopt artificial or autonomous cars on the streets before understanding, but those are the more tangible tests or last mile drone delivery. Then same thing when it comes to blockchain application or gen AI and artificial intelligence. Why am I focusing on AI now? Because the impact of AI is astronomical and we haven't even scratched the surface. They're talking the trillions and trillions of dollars and the growth that takes place. You need to make sure you create a safe environment, be more inclusive and have the right conversations. And part of creating that safe environment is of course the big question about regulation and you're right. When you're still thinking about privacy and data, privacy, then you have generative AI jump in and it changes the dynamics. How do you regulate at a time when the speed, but also even the knowledge, is not necessarily there amongst the regulatory bodies? So I'll start with you and then I'll turn to you. So I think the best way and referring to, and I know there's a question over there, so Mina if we can take that maybe after that, but I think the best way to do it, and again referring to our leadership and the expertise in AI, is to ensure that one, and I'm referring also to the professor that was one of the panelists yesterday. I apologize, I can't remember the name, but it was an amazing conversation. And he said AI is not a technology, it's an enablement. And then he also said that whoever thinks about governing and it was controversial when he said it, but now I'm connecting the door. He said whoever thinks that you can govern AI is out of his mind. And he said it in the session and I see where he's coming from because again the professor said it's a solution and it's an enablement. The best way to regulate it is to put it into buckets and buckets of industry. So how can you regulate gen AI when it comes to its involvement in mobility? How can you regulate, so you really actually focus and look at the solution as a sector rather than the word AI, because AI is going to touch us left and right and top and down. So you cannot just say I'm going to regulate AI period. You just take it bucket by bucket, create the right environment, include the right people, bring innovators, bring academia, bring the researchers, bring the right capital, whether it's equity or whether so they need, you need to create this hope for them and the right ecosystem. And like I said, the speed in the regulation needs to be there because no matter how much money you spend, how much you aggregate people, if you don't have that speed, it's game over. I mean one minor observation is that the gains from this technology as they happen in previous technologies are so significant that there's a big risk that there will be what's called regulatory capture driven by the leading firms, the big tech companies. And the UAE with its unique setup may not be at the same risk to say European countries of facing that capture, but that capture risk does exist, it's one I'm concerned with. One of the ways to counterbalance that is to create realistic alternatives, which is why things like Falcon, which is the UAE's large language model, it's open source, it's one of the most powerful and performant in the world, is really, really valuable because it creates viable alternatives to closed systems. And so when I think about this, what this means sort of more broadly, it is about ensuring there is a minimum risk of regulatory capture and there's sufficient competitive dynamics out there to keep all the participants developing this stuff honest. Okay, there's a question here if you could please identify yourself. I think this is on. Hi, my name is Rashma Sajani. I'm the founder of Girls Who Code and the CEO of Moms First. I wanted to talk and thank you, Your Excellency, this is a wonderful conversation. I wanted to talk about the importance of access and education. One of the things I learned at Girls Who Code, over a decade we taught over 600,000 girls to code, half of which were under the poverty line, half of which were girls of color. We've reached millions of girls across the world. But the problem that we were facing was that Wi-Fi for many children was not available. And so thinking about how fast AI is moving, how expensive it is to get one of these licenses to even practice with chat GPT, you know, what are we doing from what we learned in terms of coding in Wi-Fi access? And we left so many children, especially so many children in the region have been left behind. What are we doing to make sure the same thing doesn't happen with AI? That's a great question if you allow me to start. I mean, thank you so much for this question. It reminds me of an actual story that we've gone through through a program that we've launched around four years ago. It's called the One Million Arab Coders. We started with the Arab region and we said, let's come up with a program. And part of our efforts at DFF to really to raise awareness on the importance of this language of the future, if you will. So we said, okay, how about we spread the word, promote coding, then think of a way of infusing it with entrepreneurs? Because coding is not enough to build a tech company. You need the coders, and you need the developers, you need the speakers, and you need to connect everything. And you need, of course, capital. So we said, okay, we get the coders, we get the capital, and we get the academia, we get the private sector, and with the government, we can connect the dots then. So we launched that program. And to your point, I mean, it was heartbreaking when we traveled around the world. I personally went with the team to a Zaatariah refugee camp. And many of the applicants were actually from the refugee camp. And to see the skill and talent that we had, honestly, in that moment was probably one of the dearest moments for me across all my experience at DFF, just going to the camp and seeing their passion and their abilities and then thinking, okay, those are maybe 10 or 20 skillful people. Imagine the ones that we had a few examples of people actually to apply to the program, they had to actually switch locations just to get access to internet. So it shows you the importance. I know you want me to wrap up, but what we did is we got them, we flew them in, we connected them with startups as part of our efforts. And some of them right now are part of some of them are working remotely and have created successful companies hoping to push for unicorns. They're not there yet. But the idea is to really push for more of this. And I'll wrap up because I know you're not going to give me a chance to speak anymore. So I'll wrap up my side by saying the biggest message, the biggest message for AI to work is really global collaboration and understanding. And unfortunately, had to learn it the hardest way through the pandemic. It was a message that we need to unite, we need to work together. Some countries did better than others when it comes to the lockdown early, and then the vaccine rollout. And some didn't. And that size didn't matter. It was just about the mindset and how you do things. So, and you've heard that maybe in the opening remarks of his ex is Muhammad Al-Gurgawi, he sent a few major bullet points, but they were all driven by us to really collaborate and work together to create a better future. Generize the same. On the issue of inclusion. Absolutely. Well, you've brought the passion and the heart. So I'm going to bring the code in the head to the answer. If we think about where open source large language models will be within a couple of years, they're getting more and more efficient resource light. And so it's not within the bounds of probability that within two or three years, they'll have a GPT for capable large language model that will run on a medium end Android phone, even when it's not connected to the internet. And that will that will tackle this from a technical perspective. But of course, there are so many other dimensions that you need to address, which I know girls who code knows how to address, but the technology I think will work in your favor over the next couple of years. Is AI a force for good? Yes or no? Well, yes. Absolutely. Okay, great. Thank you so much for joining us.