AI and Human Development

Hello. My name is Boni de Leon. I am 23 years old, and I am a roboticist and biomaterial scientist who believes in sustainable futures. Today, I want to share three stories with you. The first story, of course, is about robots. There is a recurring pattern in this field. Engineers and scientists constantly pushing the boundaries of innovation for robots to become more human. We are making incredible progress on artificial behavior and cognition, but what about the body? We as living beings belong to circular systems. Our bodies are eventually returning to the environment, nurturing future generations of life. Can robots also perish and reintegrate into nature cycles? It turns out you can. In Japan, I designed a biodegradable robot made from potatoes with seeds hidden inside its body. When the robot's time has come, it gets buried in the soil. The body decomposes, and from the seeds it once carried, life emerges. This is the first lesson. Machines, too, can take part in the cycles of nature. The second story is about finding simple solutions to complex problems. I've been developing sustainable technologies for space. When we think about space, we tend to associate it with technology as its absolute limit, the latest scientific breakthroughs. We assume a complex problem always requires an equally complex solution, but this is not necessarily true. For example, there are insects that can be used to detect radiation because they glow under UV light, and the algae in our oceans can be cultivated in space to make concrete for habitats. Biomaterials are the building blocks of nature, reflecting its intelligence, and I am not the first person to realize this, for sure. Many of our ancestors and indigenous communities already did in the past by honoring nature in their technological practice. Many of the things they developed were so advanced and visionary. And I really believe that they can offer solutions to many of the global challenges we face today. This is the second lesson, to honor nature's intelligence and not forget the wisdom of those before us. The future of our planet may depend on this knowledge. The third story is about how human our technologies can be. I used to build robots to support people with different types of disabilities. robots to provide therapy, to help people eat. But the more time I spent with these people and the technology that they would decide for, the more I understood the importance of human care. In an era increasingly shaped by automation, I think it's essential to remember that technology does not define its own purpose. We do. This is the third lesson, to use machines as helpers of humanity, not substitutes for it. We are living in an era with achievements that were once science fiction. And it's exciting to explore how far we can go, what our limits are, what we can do. But we have somehow faith that all this technological progress will eventually save us. It is very important, more than ever, to remember where we started, in a natural world based on complex, interconnected living systems that sustain our planet. These systems have kept us alive. We must respect and protect them if we want to survive as a species. We must also remember what makes us human, our capability to love, suffer, forgive, make sacrifices, but more than anything, our capability to imagine futures and make them happen. My generation is asking today, what is the true meaning of human progress? How can we become united agents of change for the protection of our species and our planet? Which technologies should we choose not to build, even when we can? In the convening in Mexico City, we wrapped up our discussions with a small sculpture of an alebrije in the hands of each participant. Alebrijes are folk creatures made by Mexican artisans, combining parts of different animals to form a completely new creature. In some ways, AI is our alebrije, and its final form will be shaped by the collective expression of our humanity. So the question is, how do we want this alebrije to look like? A violent, destructive creature shaped by our worst traits, or a beautiful, complex being that reflects the best of us? Let's begin our conversations today with these questions in mind. Thank you very much.

Thank you so much. That was lovely. Well, thank you very much indeed, Victoria, for those opening remarks and to remind us what it means to be human. So good morning, everybody. DSG, Deputy Secretary General, Amina Mohammed, Your Eminence, Excellencies, ladies and gentlemen, I'm Zainab Badawi, and it's my great pleasure to be presiding over this very, very important conversation and discussions we're going to have at a very, very critical time. Now, you've all seen how There's been so much discussion of late about AI, and the conversations have been very much dominated by the impact on jobs, what it means for young people like Victoria and their jobs in the future, what it means for education, how we teach, how we learn. We've also heard about the possibility of rogue actors using the technology to do evil things, what kind of guardrails we should have in place, who should take responsibility for that. But what has been less discussed is the impact that AI has on human development, which is why, with her customary vision, our esteemed Deputy Secretary-General, Amina Mohammed, decided to convene some global dialogues. So she effected this in Mexico, Italy, China, and Egypt to try and cover all regions of the world, inviting people from government officials and politicians, international NGOs, domestic NGOs, community leaders, young people, also representatives of nomad and indigenous communities to just have a conversation about what AI means and what human development means in the age of AI. And the summary is all in the report that you have on your desks there before you on AI and human development. And so that's what we're going to be looking at in various aspects of this. As the mother of four children and also as president at SOAS at London University, I'm very, very very, you know, fascinated by this topic because it affects all of us one way or another, because it really goes to the essence of really what makes us human. And we've had a lot of very apocalyptic warnings about a new silicon species that's going to exterminate us all by people actually who are working in the field. So how seriously should we take all that? But to set the scene, excuse me, to set the scene, we're going to play this video just to provide a bit of a framework.

It needs to be understood that the world will not be the same again. We know that, you know, the world has seen the best and the worst of itself. And as we are progressing with the technology, if the human is not at the core and we use technology as a tool, we have a better understanding how technology is going to exist to us.

That's what makes us human, this unpredictability. AI is based on models, it's a mathematical foundation, but humans, we're based on individual experience. And that's how we transform our societies. That's the base of our culture.

The speed of AI and embodied AI is a race. It's racing to outer space speed. But our understanding and our discussion about it is not enough.

But we've had anthropologists, philosophers, artists, musicians weighing in as well. I think there's an air of optimism about the potential of AI.

But there is real concern about how AI will change or threatens to change what it means to be human.

So there we go. Just some little sound bites there. Well, let me welcome to the stage the Deputy Secretary General of the United Nations, Amina J. Mohammed. So you're waving a robotic friend there, but let a human friend welcome you to the stage.

Thank you. Thank you. Thank you, thank you so much. His Eminence Cardinal Pietro Parolin, the Secretary of State to the Holy See. My colleague and friend, Irene, who is, Doreen, who is our leader at ITU, but also on this subject. And let me just say, if it wasn't for her leadership and the voices behind the SG, we would not have had governance and science and the dialogue on AI in the UN. And so we're incredibly grateful to you as we bring the soft underbelly of AI to the discussion. Colleagues and friends, earlier this year, Pope Leo XIV wrote an encyclical about artificial intelligence and called it Magnifica Humanitas, magnificent humanity. Here is a letter about one of the most powerful technologies human beings have ever built. and he named it for us. And so I'd like to begin where Pope Leo begins, with the human being. The human brain weighs barely three pounds, yet with it, we have eradicated smallpox, mapped the human genome, and built a telescope that catches light from the first galaxies more than 13 billion years away. We've also built tools that let me in New York watch my grandchildren in Nigeria grow up. Artificial intelligence comes from the same human drive Yet almost from the start, we have spoken about it in the language of doomsday. In Mexico City, one participant compared AI to a knife. In a surgeon's hand, a knife that can save a life, in another, it can kill. But AI complicates the analogy. The knife stays in the drawer until someone picks it up. AI can be built into systems that touch millions of lives, and it can learn from us. Our science and poetry are there. So are our prejudices and inequalities. AI can both amplify at extraordinary speed and scale. So while the doomsday debate asks whether AI could destroy humanity, we ask a different question. What is AI already doing to humanity? The Secretary-General has kept the United Nations focused on what AI can do and how it should be governed. The global digital compact frames that work, the global dialogue on AI governance and the independent scientific panel, our AI on AI taking it forward. And the new UN system data commons puts trusted data into people's hands. Our convenings are something different. What does AI mean for human development itself? We took that question to Cairo, to Mexico, to Sian. Rome, and to a huge group of young people in India. A priest sat beside a robotics engineer, a philosopher beside a Nobel Peace Prize laureate, anthropologists, artists, indigenous leaders, and young people joined the conversation. Across four very different conversations, similar threads kept coming through. The first was a question of human capability, smarter machines must not make for less capable people. We heard concerns about skills weakening through disuse or never developing because the system steps in first. Can people still exercise their own judgment, agency? Can they understand and challenge the systems that are shaping their lives? The second was whose knowledge these systems are built on. AI is not culturally neutral. The systems used across the world still reflect a narrow slice of humanity. If our languages and histories are absent, whole societies risk seeing the world through someone else's frame. A girl in Kano in my country, Nigeria, should not have to learn about the world from a system that has never heard her language. The third was responsibility. When we say a machine decided, we lose sight of the people who designed it, chose where to use it, and set its limits. A machine cannot be held to account. People can and must. The fourth was more difficult. What should remain distinctly human? The better AI becomes at imitating human behavior, the more carefully we need to decide where imitation is enough and where the human being still matters. The fifth, the report warns that many of these choices will be made in months. rather than years. And we heard that this morning when we spoke to what philanthropy could do to support AI. This is not a two-year process, it's a two-week process to get support behind the work that has been done in the field. Finally, across all four conversations, one question kept returning, who gets to decide? In Cairo, Graca Machado stopped the room when she questioned a single word. We keep saying that we need to decide what AI is for. And she asked, which we are we referring to? If this technology is going to shape the human family in all of its diversity, everyone must have a voice in shaping it. As the Secretary General told the General Assembly yesterday, the greatest power is not the power to dominate. It is the power to build a future that works for everyone. Friends, ladies and gentlemen, we do not yet have every answer. Those questions continue in this room and others today. And from here, we take them back out to the world. In 2027, our resident coordinators and country teams of the UN will take these questions to country-led and community-based conversations. We want to hear directly from people about how AI is changing their lives, what they want from it, and those conversations will inform the 2027 SDG Summit. But make no mistake, we have to give a certain literacy. We have to educate people, not to assume that because you can pick up a phone and you're using AI, you actually understand the full ramifications and what it is about. It's our duty that part of the education system is to bring us to the fore of the knowledge that we need to have to use the agency for what we want. At the center of the sustainable development goals is one promise. It's about the whole of our human family, to leave no one behind. And as we look to 2030 and beyond, that promise must guide us in the age of AI too. Whatever AI becomes, the measure of progress must still be what it makes possible for the human being. And so I come back to the two words from Pope Leo, magnificent humanity. our magnificence is not simply that we built AI it must be whether we are wise enough to decide what it is for disciplined enough and responsible enough to set the limits and generous enough to expand possibility for the entire human family thank you thank.

You so much do you want to stay on stage for the Thank you so much indeed, DSG Amina Mohammed. I'd like to invite to the stage the Secretary of State of the Holy See, His Eminence Cardinal Pietro Parolin. And of course, as both the DSG and as you saw in that film, it's just a few months after that, the Pope's encyclical Magnifica Humanitas. And then that was followed by discussions about AI and human dignity and so on at the Pontifical Gregorian University in Rome. And you're going to share with us your Eminence, some of your thinking. Thank you. Thank you. Thank you so much.

Madame Deputy Secretary General, dear participants, the Holy See wishes to sincerely thank Deputy Secretary General Amina Mohammed and the United Nations Office for Partnerships for convening this high-level meeting on artificial intelligence and human development. This meeting brings to New York a conversation that has traveled from Cairo to Mexico City and from Xi'an to Rome at the Pontifical Gregorian University. The fact that these discussions have been hosted in such different contexts, each with its own cultural or religious background, demonstrates the importance of basing the debate on AI on universal values in order to ensure that its development and use conform to the essential purpose of integral human development. The Holy See is firmly committed to this conversation on AI, where the dignity of the human person is truly at stake. The current age of change shows with ever greater evidence that the developments in AI are opening up opportunities that until recently seemed unimaginable. These developments can indeed contribute to human development and the common good. This is already the case in fields such as medicine, agriculture, education, and culture. However, it is important that they are never separated from the discernment related to the anthropological vision that guides them and the goals that they pursue. As Pope Leo XIV stresses, if I'm quoting, "If technological development advances without a corresponding ethical and social progress, the result may be an increase in means without a growth in humanity." For instance, the world of work offers a relevant point of reference. While technology's potential to spare human beings from repetitive or burdensome tasks should be acknowledged, it cannot come at the expense of the irreplaceable role of the human person. This is particularly pertinent at a time when technologies are advancing in unpredictable ways, making it difficult to fully assess their long-term impact on human dignity and the common good. Thus, for the discernment to be fruitful, it is necessary to embrace the concept of limits, which may initially seem contrary to the very idea of progress. As Pope Leo XIV insists, calling for prudence, rigorous evaluation, and even at times, a slower pace in adopting AI, does not mean opposing progress. Instead, it is an exercise of responsible care for the human family. This need is all the more urgent given the frequent imbalance between the speed of technological growth and the slower development of awareness, norms, safeguards, and institutions capable of governing its effects. It is not enough to invoke ethics in the abstract. Robust legal frameworks, independent oversight, inform users in the political system that does not abdicate its responsibility are required. The establishment of the global dialogue on AI governance reflects the will of the international community to act responsibly by discussing on an annual basis the social, economic, ethical, cultural, linguistic, and technical implications of AI, also based on evidence brought in by the scientific panel, with the view to developing international cooperation and best practices. Madam Deputy Secretary General, distinguished participants. In a context where digital networks and developments in AI are increasingly bound to the globalized economy, amplifying the effects of decisions made, there is a risk that those holding greater technical and economic resources may impose their own particular vision of reality. It is therefore incumbent upon the international community to govern and orient this age of change so that the promotion of God-given human dignity of every person, rather than technical or economic considerations, remains at the core of every decision that shapes technological progress in the years to come. I thank you for your listening.

Yes, you can shake his hand. Thank you so much, Andy. Thank you. Thank you so much, Your Eminence. Thank you. His Eminence Cardinal Pietro Parolin reminding us about God-given human dignity. And for those, of course, who are believers in God, it's a reminder of how belief is an instinct, which is not something that can be proved, can it, by machines. So a very, very important message there. I'd like to invite to the stage, well, in the order they're coming on, Gillian Tett, who's a very distinguished journalist, somebody I've admired for many years, who's from the Financial Times, a columnist. And she's also on the board, the editorial board of the Financial Times. Gillian, why don't you sit here, yeah? And then we have also Doreen Bogdan-Martin, Secretary General, please do come up with the, oh, you're not Doreen. Oh, okay, that's fine, I'm flexible. Brad Smith, who is President of Microsoft and is also I'm here. You'll be another chair.

Am I here or there?

We're getting this.

Brad Smith.

Am I to the left of Brad?

I'm still here. Yeah, that's great, Mike. That's fine. They're all the same chairs. They're all equally comfortable. Brad Smith, president of Microsoft, also the Secretary General's advocate on sustainable development goals. Doreen Bogdan Martin, who is Secretary General of the ITU, which is the International Telecommunications Union. Welcome to you all. So, You've heard what's being discussed. We're going to be talking about AI and global human development. Just looking at my clock here, we've got about 17 or 18 minutes or so. Gillian, although you work for the Financial Times, you are actually a PhD holder in anthropology. So I'd like you to put your anthropology hat on and discard the FT one just for the moment. And just to say, you know, it's not all doom and gloom as both the DSG and the Cardinal said. There is, you know, optimism around this whole issue of AI.

Absolutely. Well, I'm speaking both as a journalist and as an anthropologist and someone who's now at Cambridge University. And I'll say three quick things. One, we are talking about the wrong A. I think we should rename it not artificial intelligence, not even superior intelligence, although that's what the president of the US said yesterday, but augmented intelligence, because we need artificial intelligence plus anthropology intelligence to create augmented intelligence, i.e. a tool that makes humans more clever. And in fact, in this fascinating survey from the convenings, in fact, there's a significant chunk of the participants who would say also it's about augmentation. Secondly, The global south has a lot to teach the developed world about how to imagine the relationship between humans and AI.

Give her an applause for that, I think.

Because although Silicon Valley dominates the debate right now, that is not the only framing of how humans imagine the wider universe and intelligent machines at all. And I was reminded of that because Cambridge University convened global regulators last week to talk about regulatory tools to oversee AI, and I have to say that it was a team from the global south who came up with some of the most creative and effective solutions of all. Groups like the UAE, Rwanda, India are doing some remarkable things, to cite just a few. And the third point is very obvious, need guardrails. We can borrow from the world of finance, from nuclear weapons, from bioterrorism spheres to look at how we create guardrails. And talking about how we create guardrails right now is absolutely critical to build the trust. Because at the end of the day, the dangers from AI don't so much come from the machines themselves going mad, they come from humans using the machines to do terrible things, non-state actors in particular. And that goes back to the point. It's really about recognizing this is about augmented intelligence for good and bad.

So let me go to you, Brad. Do you agree that, I mean, that's a relatively sanguine look at the whole issue because there are those who believe that AI is already challenging what it means to be human. You know, there are surveys that show young people are losing their ability to empathize because they've got relationships with chatbots and all the rest of it. And so the very essence of human collaboration, cooperation, knowledge, and so on is being challenged.

Well, I would agree with most or perhaps all of what Gillian said, but let me pull on that thread on human development and let's talk about kids, let's talk about students. I think at its core, we need to use AI as a tool, but keep it in its proper place. And just the most thought-provoking aspect that I've encountered in the last month was just being here in New York two weeks ago. We at Microsoft entered into a binding contract with the second largest teachers union in the United States that will become part of a binding contract with every school district in the country. And I think to some degree, it's a role model for others to think about around the world for precisely the reason you just mentioned. We agreed that we would contractually prohibit ourselves from providing students with any AI tool that would foster emotional attachment or dependence, that we would not provide any AI tool that would engage a student in a learning task or in the use of the product longer than the learning task required. That we would not provide an AI tool that would simulate friendship. And we would not provide any AI tool that would encourage students to divulge sensitive information about themselves. What this says at its heart is that what is most important for the development of children, students in schools, is to learn from their teacher and develop relationships with their classmates. use AI as a learning tool, but don't substitute for that. And I think that says all you need to know about human development. Let's have humans develop by continuing to connect with other humans. I think that's at the core of Pope Leo's encyclical as well.

So, Maureen, I mean, you're from the International Telecommunications Union, so communication collaboration, cooperation. This is how human society has developed over the centuries. Do you feel that when we're talking about AI and human development, that that is something that's a bit under assault?

Well, let me start by commending the DSG for her vision and her leadership on this issue. I think it's critical, as she mentioned, and as the video highlighted, we have to keep humans in the center. And I just wanted to also mention, I thought it was quite striking yesterday in the General Assembly, we had 30 leaders speaking, plus the Secretary General and the President of the General Assembly. 75% of those interventions had a big focus on artificial intelligence, something we would never have imagined, I think, even last year. And so we ask why, and I think it's kind of obvious why, because AI is critical to development, it's critical to societies, to economies, to security, and to peace. But at the same time, and this comes back to your point about connectivity, we're living in a world where AI is like a mirror. And so AI basically as a mirror reflects what it is shown and we think about what it is shown well we have 2/3 of Africa that's never ever been connected to the internet we have over 70% of the population that has never used uh generative AI we have electricity challenges as Brad speaks about often And we have this incredibly powerful tool that is largely focused on the English language, which most of the world doesn't actually speak. And then we have these huge issues around safety that was also highlighted in so many of the discussions yesterday. And I think if we don't tackle those safety issues, we're not gonna have trust. If we can't find ways to tackle the trust issues, we're not going to have adoption. And then just perhaps very quickly, I think at the same time, we have to realize, this comes back to some of the examples in terms of the global south, so much is happening. There's so many great things out there that's actually helping augment our humanity, I think, when it comes to examples in education. We had this great case that was shown a couple months ago about something created in Nigeria to help illiterate students. We have incredible advances in healthcare, even for low resourced communities, incredible advances in agriculture, also for low resource communities. So The solutions are there. The global south in many ways is leading on some of those solutions that tackle community challenges. But we have so many other issues that we have to address to help it reach its potential.

I mean, you brought up the point which the DSG mentioned about the large language models so-called. And I mean, I'm struck by the fact that Welsh features more greatly in large language models spoken by, and important, I don't wanna offend any Welsh speakers, 'cause I, by the way, this is all going out online as well. Hello, Welsh speakers online. But they vastly outnumber Swahili, you know, references and so on. And that goes to the point of the, who is the we and who is this augmented intelligence as you'd rather call it or artificial intelligence who is it representing Swahili spoken by millions and millions of people in in Africa and yet it's not you know this kind of knowledge base that you say the South has Global South has got a lot to teach the global North the rest of the world is not really represented in the current system.

Well I'd like to say that in some of the practical applications of AI. In fact, parts of the global South may end up moving ahead of the West. To give one tiny example, the UAE has created a challenge for the government whereby they're hoping to put 50% of government services, they're hoping will be performed by generative AI platforms within two years. I don't know any Western developed world government that's actually pleased to do that. To give other examples, you look at something like the Indian IPU stack and the astonishing success of that. and look at how an entire society can potentially get access to these tools in really quite creative ways. Again, the West has yet to get digital IDs rolled out on a massive scale to do that in large countries.

But I would say your point about language is still valid. And look, the UAE is leading the world, literally leading the world in the percentage of the population using AI. And it's not just using it, but there's just extraordinary progress there in using AI to improve healthcare, to invest in education. Of course, there's not too many people in the UAE who speak Swahili or the 47 other country languages that are spoken by significant populations across Africa. So the problem you described needs to be solved. It actually, frankly, in the scheme of things, is one of the easier problems, I believe, to solve with the right investment. We, Google, others have been working on this for a couple of years. The Gates Foundation brought together 60 partners, including the UN and the UK government on Monday. What do we need to do? Well, first, you've got to get more content in these languages, often digitized. so that they can get them on the internet. The second thing you need to do, which actually isn't that complicated, is you just need to capture a few hundred or more hours of audio conversation so that then you can train AI models in the speech-to-text, text-to-speech, and other capabilities that are needed. Those are relatively easy aspects. The part that actually is harder, but in some ways is even more important in terms of the long term, is to train more data scientists. I'll say especially across Africa. Because for every 14 data scientists you'll meet in Europe on a per capita basis, you'll meet one in Africa. So we need to invest in education, and especially in that field. That's part of this solution.

Yeah, I mean, you're absolutely right. People often talk about the mobile phone technology which allowed Africa to leapfrog, but when it comes to this issue, you do need the trained specialists and the data centers as well. You brought up the issue of trust. So until what Brad Smith is describing actually happens, there is going to be an issue of trust in terms of AI and human development for large parts of the world.

No, exactly. And also, I mean, congrats, Brad, on the effort you just mentioned with the Gates Foundation, 'cause I think that that's really critical when it comes to language representation. But obviously we need to focus on this trust issue, and it's also linked to inclusion. and to the DSG's points, it's really critical that we have all voices in this conversation when we think about AI today, when we think about AI in the future. And we have to have different voices, different age groups that are part of this conversation. And I think also linked to the adoption and the trust piece, we have to think about the impacts on education. And so something coming back to the UAE example. So the UAE has actually set up a university, an AI university. If you take, for example, and I think we're gonna hear from her this morning, Minister Paula from Rwanda. So Rwanda has taken a decision that all teachers in high school and elementary school have to be AI literate. That's also something we're not doing here in the United States or in developed countries. So we're seeing lots of innovation when it comes to education, lots of ways developing countries want to ensure that their younger generations are AI literate.

So when you say AI literate, the teachers, I mean, they've also got to be quite canny at checking what is just the copy and paste that the students are doing. But it is transforming the way teachers teach. because you know AI is used as a tool so we have to accept that it's used as a tool in education by students but you're saying they need to be taught how to use it in a judicious way but do you think it's going to transform the way we look at education in terms of the kind of exam system a lot of Countries now are saying, governments are saying, we need to have more oral exams. There's a huge disparity between the coursework that's submitted by students, and yet when it comes to exams and oral exams, they don't do as well because, of course, they don't have the AI at their disposal.

It's entirely possible we'll end up with human-to-human interaction and the teaching system being the new luxury good. And we need to make sure that that is not just retained amongst a small group of elites, But I'd say the key issue right now is that alongside teaching STEM skills, we need to teach critical thinking. And we need to keep both in focus. And I oversee King's College in Cambridge, which was the home of Alan Turing. People sometimes say we need to have Turing-based education, lots of AI-based education. We actually need Turing plus. Turing plus critical thinking. And that's going to be the crucial ingredients to keep together, but not just in the wealthy countries in the world, right across the global education system as well.

You all agree with that, because there's a huge emphasis on STEM subjects all over the world. And you've got to accept that the arts, humanities, the social sciences, those subjects which teach us the critical thinking, not that science doesn't, in a way, mathematical, like mathematics, I suppose, does, but understanding different cultures, not othering, people who seem different from you and thereby promoting what UNESCO often refers to as a culture of peace. It's very important if we want to build and develop the kind of world we want, that we have those kind of subjects still, the ethicists, you know, the philosophers and all the rest of it.

And the anthropologists?

And the anthropologists. I always count in more arts, humanities, social sciences, absolutely.

Look, yes, the world needs, frankly, more critical thinking than perhaps AI. But let me just also add this because you started this thread by pulling on safety. The world is going to an ethics class today and all of you and all of us are part of it. And it's about fundamental ethical values because that in part is what this safety issue is about. Imagine for a moment that there's only two big companies in the world that produce commercial airplanes, it's not that hard to imagine. Now for each of you, because most of you are younger than me, imagine that you're the CEO of one of those two companies. And imagine that somebody comes into your office and says, I know we're in a race to get out the next airplane first, but we're worried that the airplane will crash and people will die. You're the CEO. This is your ethical test. Would you say, I need to win this race? I won't slow down unless my competitor does. I won't slow down unless the whole world agrees to slow down. Or would you say I am going to put the safety of my passengers and the safety of the public ahead of my winning that race? That is the fundamental question for people who work in the tech sector today. The journalists of the world are treating this as a complicated question. Oh, how can people slow down unless the US and China come to a grand treaty? How can they slow down unless the United Nations adopts a resolution requiring everyone to slow down? Is that really a complicated question? Or should we all as individuals just remind people there are fundamental ethical values. This is at the heart of critical thinking. This technology may be complicated, but this of all things is not.

Thank you. A very quick question to all of you just to wrap up. Rod. You're an advocate for the Sustainable Development Goals. And of course we know that it was the DSG Amina Mohammed who spearheaded these. We heard her and also his eminence, the Cardinal, speak about the importance of putting human dignity at the heart of the SDGs and so on. So very quickly, how far do you think AI can actually promote this sense of solidarity and human dignity within the framework of the SDGs? I'll start with.

So the SDGs are way off track, as we know. And I think it's important to remember that we're having this conversation in a group of kind of, we're all connected digitally. And when we think about the SDGs, we have done extensive research with our friends at UNDP. and demonstrated that if you invest in digital technologies and put AI on top of it, you'll get close to about accelerating progress on 80% of the SDG targets. Because whether it's education, healthcare, agriculture, whatever, digital technologies and AI can help unleash incredible opportunities.

Thank you, Julian.

I would say there are two ways. You are either good on the route of the pitchfork or the Pope. Pitchfork is basically talking about the fact that if we do not focus on the SDGs, we're going to end up with a world of war. ripped apart by war, terror, hunger, and threats that at the end of the day will come back even to the rich. Just think of what happened with COVID and the pandemic. What better example do you need to find that actually if you ignore the plight of the world, even the rich will be infected, and they should be concerned about that too. So you can go down the pitchfork route. Or you can do what the Pope, his Holy See, has done with extraordinary skill and inserted himself into this conversation, inserted the church with its values into this conversation and say, there is something higher that we can all as humanity essentially aim for. And it's not just about the Catholic Church. I was at an event last week with Hindus talking about what Hinduism has to add to the debate about AI. All of the world's religions and ethical traditions have really valuable perspectives right now. And so I would prefer to go down the route that's been illuminated by the Pope and talk about our global humanity rather than just focus on the threat of the pitchforks. Because if we don't do that, those pitchforks are coming.

Thank you.

Just briefly, I would say the 17 SDGs remain of paramount importance. And as we get to 2030, no doubt we'll have to devise successors to it. AI, I think, is important in three ways. First, we have to spread access to AI so that every part of the world can make use of it. Second, we need to use the benefits that AI can offer because you take any one of those 17, look at the problem, there is an AI tool that can help solve it. But then third, we need to keep it under human control. We need to keep it safe. If we don't, then we have bigger problems than we have today.

Thank you very much indeed. Thank you to our first panel.

Thank you.

So you've heard um our two keynote speakers in our first panel you can see on your desks so now you've had time for the ideas to percolate I want you to answer three questions which are going to come on the board Ella and James please do come onto the stage while um you while it's being um assembled yes thank you could I have my back. Yeah, so could you put the three questions up on the, uh, on the board, please? So you've got a red and green card on your table. If you agree with this question, does AI respect human dignity? Raise your green card. Can I have my papers back? Does AI respect human dignity? Raise your green card. If you disagree, raise your red. Oh, my goodness. I think I'm seeing a flurry of red. Gosh. That's, I think, overwhelming. Right. Okay, next question. Raise your card. Can humanity remain in control of the future it is creating? Raise your card green if you agree and red if you disagree. Can humanity take me, you know? Oh, more green. So you're optimistic there. And then the third one is, does technological progress necessarily mean human progress? AI, of course, is a subset of technology. AI and technology are not necessarily the same thing, but does technological progress necessarily mean human progress? Wow, a lot of you disagree. bunch of pessimists, aren't you? But you do believe that humanity can actually do something. Will it is a different question. Do you think it will? What about that? You all think it will. Yeah, that they can actually put. So you're optimistic there on the fact that humans will retain agency and use AI and technology as a tool, but you're not sure about it respecting human dignity. interesting, something for our next panel to consider. Ella Alshamrany, please come to the stage. Ella is a British paleoanthropologist and television presenter. And James Manyika, President of Research Labs, Technology and Society, Google, Alphabet. Please take your seats. Thank you. So this time, we're going to talk about AI and the evolution of human civilization. So, just looking at the time. So, Ella, you've studied genetics or a paleoanthropologist and so on. So what's new and different this time? I mean, humans have always developed. Well, fire, we started using fire about what, a million years or more than a million years ago. And I mean, fire is useful, isn't it? Humans have used it, but if it runs out of control and runs amok, it's very, very destructive. So It is, I mean, you know, is this just a continuation of the kind of inventions and discoveries that humans have made and that in a sense, we will learn to harness the power of fire AI and use it in a controlled manner without it running amok.

So I'd say just generally speaking, you've got to be a bit cautious about asking a historian to predict the future. Having said that, I do think that there are trends that we see. We have seen a trend, in my opinion, since the beginning of agriculture. So, quick history lesson, for most of our time on this planet as a species, we were hunter-gatherers. We did not have a permanent abode. And then somebody decided to plant the seed in the ground and we had agriculture. And since then, the technological changes that have happened to our species have been so considerable and in my opinion have been the greatest trade off we have ever made. We have, as a result, Things that are beautiful, like the Sistine Chapel. Things that take your breath away. Books that engage us, make us think. The kinds of things that we would not have had. Also, there's more of us on the planet, which, let's be honest, most of us would not be here if we didn't have agriculture. However, the flip side is that we're probably not evolved for the world that we're in anymore. And so I would say AI is just the latest manifestation of this. It is just the latest iteration of a deal that we made starting 10,000 years ago. We know how this goes. It's a trade off. There will be benefits. There will be negative sides of it. And we've got to navigate both of those, I would say.

James, how do you see this question of AI and human civilization?

Well, actually, Zena, building on what Ali was describing, I was struck by, in his 2025 Nobel lecture, the economic historian Joel Mok made this point that if you look at the long arc of history, technological progress is always driven humanity's progress and increased social welfare throughout history. I think in that sense, it's a continuation of that trend. But I think it's worth thinking about this moment when it comes to AI with its particular characteristics. I think on the one hand, the analogy that's often made is obviously is the idea of AI as a general purpose technology. Which is the case, I think it probably has the strongest claim along with electricity and other things to be a general purpose technology in the sense that you can imagine it being used across a wide range of human endeavors to enable them and power them. And so that analogy of the general purpose technology typically then makes people make comparisons with industrial revolution, which I think is appropriate. But I think there's a second set of characteristics that are also important to keep in mind. This is the idea of AI as an invention for making inventions, which is a different set of characteristics. And I think there we should imagine the kind of analogies we should be applying there is what happened to the Enlightenment and scientific enlightenment, where we created new tools, new instruments to see the world, microscopes, telescopes, and so forth. We created new methods. to make discoveries, the scientific method itself. So I think it's these two characteristics of the general purpose technology and an invention for making inventions that make this a remarkable technology full of possibilities. But I think as the previous panel pointed out, the question isn't to create this technology for its own sake. The question is to ask the, you know, think about why are we doing this, what is the why of AI, why do we wanna do this? I'm struck by my friend and colleague, Demis Hassabis, who said, Demis Hassabis, another laureate himself, who founded DeepMind, I think put this very well at the time as a mission for DeepMind. He said, we want to solve intelligence and use it to solve everything else. And I think the everything else is in service of what humanity would like to do and should aspire to do over the long arc of history.

Ella, it's interesting to hear what James says, and he talked about the tools and the industrial revolution and the machinery and so on. And people often talked about Johannes Gutenberg, the man who invented the mass printing press in the 1440s and how that transformed the world and displaced jobs and so on. But to what extent is this different? Because what makes us human as you all know, is our power of speech. Of course, other animals communicate with one another, but we uniquely have the power of language, words, writing, and reading, and so on. And what AI can do is speak like us. We have gifted to a silicon species when they become robots and so on, something which has been uniquely human. And I think it's that kind of threat which people are thinking actually is AI representing something which is quite different that we've not seen before, you know, the tools of the industrial revolution, machine and AI and so on.

I don't know. I think the truth is the invention of writing had Plato worried You agree, right? So for those of you who don't know, Plato was concerned, to paraphrase, that the invention of writing would basically make people dumber because they would no longer have to rely on their memory. I think with every new technology, it is very natural for some people to be resistant and for some people to be fearful because it will inevitably lead to some kind of a social, a change in the social fabric. I think it is just the latest iteration of that. It's quite a potentially quite a huge variation on what we've seen before, but it is just one of the same. I would say, you know, I have people coming to me asking me, are our brains going to get smaller because of AI? And, you know, I think this is so intriguing. First of all, you know, genetically, that's not going to happen because just, I could go into the biology of it, but that's just not going to happen. But I think the bigger concern for me is how do we look at this new technology in a world where we are now surrounded by technology evolving at such a fast pace and acknowledge that we are not evolved for the world that we're in? So when we talk about humanity, part of the thing that made us who we are is cooperation, right? We were not designed for a loneliness epidemic. We were designed for tribes of, we used to live in tribes of a few dozen people. You knew everybody by name, mostly because they were your dad's relatives. And that's how our mental health, our physical health, so many things about us are designed for that world. We no longer live in that world, most of us. Most of us have abandoned that world a long time ago. And so for me, the question is, how do we remember the things that make us human and look at this new iteration of technology and say, right, how do we protect some of this stuff? I think it's really, really complicated. I think we should acknowledge that a simple, a simplistic, this is a brilliant technology, this is awful technology. It's just not in keeping, in my opinion, with the last 10,000 years.

It's interesting. Ella raises, James, an interesting point. And of course, evolutionary biologists, and you are also that, tell us, of course, evolution is complete. The human brain, as you say, is not evolving anymore. But does that It doesn't mean to say, though, that we will not cease to use the left hemisphere of our brain less. I mean, you know that, you know, psychiatrists and, you know, experts in this field, neuroscientists say there's the right hemisphere and the left hemisphere of the brain. And the right hemisphere is much more like a river flowing, meandering, paying attention to the world in a different way, seeing things as not fixed and so on, creativity, imagination, intuition, insight. to experience, that's our right side of the brain, which is, I think, more important than the left side, which is much more the hard facts, you know, like stones, the solid things and so on. Of course, humans need both in order to figure out the world. But could it be that because now people don't have to use their memory, they'll just press a button and find out some information, that that left side of the brain perhaps won't be used as as much because we are relying on the machine so you know although the brain stays the same the way we use it may be different.

I'm not sure I agree. I think, first of all, we don't fully understand the brain, first of all. It's still a long way to go. In fact, just recently, my colleagues and I, finally, there's been a long process of trying to understand the structure of the brain, the wiring diagram of the brain. It's a field called connectomics. So we still don't quite understand the brain. So I think it's too early to place limitations on what the brain is capable of at this stage. So I think there's still more we're going to discover. But I think in this regard, I think, and we've gone this before with calculators and other things, and Ali pointed to even writing. But I think what's exciting to me about the possibilities, particularly of this technology, as I said, as an invention for making inventions, is the questions we can now ask, the things we can now aspire to solve, the kinds of things that we can work on that advance humanity. And I think in many regards, I mean, one example, and also what it means for all of us, One example is in our world, in my case, in our case, is the work we've done on AlphaFold, which is the breakthrough insight there was to understand the structure of proteins, all of them, not just one or two or three. We only understood about 100,000 proteins, 170,000 proteins before that. Now we know all of them. That's an incredible breakthrough. But the reason that example is a useful one is not only do we make a breakthrough, which opens up opinions of research, we also opened it up to everybody. So up until now, as of last week, I think we now have 4 million researchers in over 190 countries using it. In fact, I'm actually particularly proud of the fact that 32,000 of them are in Africa. In fact, I came across the other day a young researcher from Zimbabwe, Tinashe Muyambo, who's using AlphaFold. And many of these researchers in many more places are now working on diseases, many of which are neglected diseases.

I don't want to duplicate what we were talking about in the human development panel, because this is about human civilization in a sense. You say that it's not going an impact perhaps on, you know, what we think, but values are very, very important, human values. And is there perhaps going to be an assault on those human values? You said that they're already disappearing to some extent, the way we lived in a very collaborative way in the past, where you don't bring up your own child, you've got the whole community helping you do that. But you know, as people increasingly talk to their chatbots and they agree with everything they say and so on, the ability to empathize wisdom, trust, all these things which make us human, compassion, passion. Can machines really be passionate, compassionate? Can they show wisdom or whatever? So are we going to perhaps find that human civilization, the way we develop, is under a sort as these human values become devalued in the age of the machine? both of you.

I'll just go back to what you were saying. I just want to add one thing about the brain. It's that even if we, our generation, the generations that come after us, start using one part of our brain because AI is, let's go with that thought process, truthfully, because there will probably be no genetic change in It is all reversible. We could set an education standard or an education system to think differently. So I think we're actually, I don't want to say we're in a good place, but I don't know if it will have that kind of an impact that is irreversible. I think that's really, really important and hopeful. And maybe we go to the next question.

Yeah, I know, I was going to say, but if you think about the way we interact with machines now, it's human curiosity. curiosity that drives you into asking AI a question. It comes from human curiosity. You mentioned the Nobel laureates. Their work, their breakthroughs come through their curiosity. They want to know something. So that's what I'm asking you is to, you know, if you can find out a lot of information at the press of a button, are these things which have developed us, developed our civilizations, going to perhaps become much rarer in the age of AI.

Yeah, I don't think so. I think this will allow us to ask new questions that we could not even have had the chance to even ask. So for example, I think a lot now about all the new kind of what if questions we can now ask, which we couldn't have asked before. So right now we can ask the question, What if every disease on Earth was diagnosable, understandable, predictable, far in advance? I don't think we could have asked that question before. What if every natural disaster or natural event on Earth could be predicted so that no human could be surprised by that natural event? I think you can go on and on and on. What if every human could communicate and interact with every other human on Earth without the boundaries of language and access and all of those things? These are fundamentally new questions that have the potential to advance human development. So we can now ask those questions. The good news is that I think we're starting to make progress on all of them, actually. So, for example, in our case, we're working on all the diseases, trying to predict and understand and diagnose them, breast cancer, colorectal cancer, and all the cancers, tuberculosis, which affects 40% of the world's population goes undiagnosed because they live in places and communities where we can't do that. Think about natural disasters. We're working down a list of everything from floods to monsoons to cyclones, even earthquakes. Recently in the recent Venezuela earthquake, we're able to send alerts.

I mean, that's even that's the benefits of AI.

Yeah, these are questions. What if people could talk to each other? If you want to have interactions across people on Earth, you'd have to do.

7,000 languages, we have a microchip that puts you and you can speak fluent whatever.

Right. But so I think the possibility.

Of a lot of time, won't it? Studying all those nouns, declensions, yes, final words on this.

I think it was kind of, I kind of wanted to ask you the question a bit more on this because I think what you were saying about concerns about technology generally, and AI just being the latest iteration of that, as great as it is that we could end up with a universal translator, I think I'm more concerned that AI, like social media, like the invention of cities, has made us actually more insular and less community orientated. And our superpower, in my opinion, as a species, is our strength of cooperation and our strength of human interaction. We are the most cooperative species that's ever existed. I guess My question is, how do we maintain all those positives while acknowledging that there is this other side to it?

And I agree with that. I think in some ways, one of the things I'm excited about when it comes to human development is that, in fact, this technology might actually allow us to discover the deeply human things about us, interactions, real-world relationships, which I agree with, society, community. even emotional intelligence. I think a lot of our effort has been, I mean, imagine if we all spent all our time learning just multiplication tables and doing those kinds of things. But now we can actually open ourselves up to relationships, to interactions, things in the real world, asking us new questions. I think it frees us up.

Using the creative.

Oh, absolutely. Absolutely. I think these are tools that should be in service of human development, human progress. And I think we should think.

What Ella is saying also, though, if we know what the machines can do better than we are, it raises the question of what can we do that the machines should never be able to do or can't do and these are the values the collaboration communication the things that make us human you know how do you maintain the human how do you maintain the age of AI that is the key.

And I fully agree with that and I think that's a fundamental question we now have the luxury in some ways to tackle and deeply understand as opposed to spending time on things that sure our tools can support us and help us do.

I'm still looking forward to that chip that means I don't have to improve my French because I am desperately to do that, if I could just insert it. James and Ella, thank you very much indeed. Thank you. Thank you. We have our next... Thank you, James. We have our next panel, if I could invite our panelists, Paula Ingabire from Rwanda, Enrique Ochoa and David Sengay from Sierra Leone. And you can see who they are there. David Sengay, Chief Minister of Sierra Leone, Paula Ngabire, Minister of ICT and Innovation from Rwanda, and Enrique Ochoa, Undersecretary for Multilateral Affairs. Don't move my table.

I'll come this.

Way. Thank you. So I've got everybody's names on them. One more chair here, maybe. Thank you. It's interesting, isn't it? You can see how there are a lot of questions that are raised.

Perfect.

So welcome to you all. So what we're talking about here is the future we choose. You saw how our earlier panelists and our speakers are talking about, you know, the all the questions that this raises and so let me start with you Paula because you took part in the convening in Cairo didn't you about how so just tell us how have the how has this whole series of AI and human developed development helped shape your thinking because I know of course Rwanda is one of those countries in Africa that is really streets ahead in its digital transformation agenda.

Thank you. I think in a way, when the AI and human development series of conversations started, what we had already had in place as Rwanda was our national AI policy and strategy. And at the heart of it was how do we center humans? Because it's not just the sake of implementing AI as a new technology on the block, but it's really how is it going to serve our nurses better, our teachers better, our farmers better, our people better. And what has helped over time is because of the focus of the AI and human development series is consistently centering that conversation and ensuring that as we deploy this technology, we're also thinking about the human at the center of how we deploy this technology. And so in many ways, resonates with the strategy and, you know, execution programs that we have in place. But at the same time, I think at a global perspective, it allows us, because as you think about AI, there's so many dependencies on big tech companies, on the development partners, so ensuring that all of these, you know, partners and partnerships can come together to reflect on how, as we deploy AI, we're making sure that the human remains with agency and choice over time, but also at the same time, how we build their capability to make that choice.

Interesting. Enrique Ochoa, so, I mean, how are you ensuring that Mexico is really incorporating its rich, diverse culture and, you know, indigenous languages and so on, because the way AI is being developed, I mean, it really is very much a technocracy, isn't it? That's based in just, you know, tiny bit of the world.

Yeah, no, thank you. I mean, coming from a country with more than 60 ethnic groups with more than 60 indigenous languages. Of course, incorporating that diversity is very important. And our perspective in general is that, I mean, AI should support that rather than divide us. And this is part of the richness of what we think we should incorporate into AI. And therefore, we developed this in early during this year in In January, we published this Chapultepec Principles. And the Chapultepec Principles are basically 10 principles. And among them, of course, human rights is one very basic principle of artificial intelligence. The other one is, of course, incorporating cultural diversity into AI. Whatever is related to AI has to take into account cultural diversity. Another point is transparency. Transparency is incredibly important when we're talking about AI and inclusive participation. But also something, a point that this Chapultepec principles include is the issue of education. Because, I mean, we celebrated in Mexico, in Chiapas, in San Juan Chamula recently, the World Literacy Day. And in a country where, I mean, Unfortunately, our literacy rate is below 4% right now. But when you're thinking about AI, this is also something that you have to consider because for those less than 4% of people that are not able to read, talking about AI is something completely abstract and they will be completely left aside. So education is also one of the principles that we have to address. And this is, I think, something that happens all around the world, particularly in the developing world where literacy is still a problem. And then another dimension that is very important for us is that all decisions that are made through artificial intelligence have to have some kind of human supervision. And this is just to also to share with you that for Mexico, this is very important. And in the General Assembly, we recently proposed a resolution which was adopted unfortunately by a vote and didn't join consensus on having a while. Nuclear weapons exist. There has to be a human supervision. And this is something that of course, like when you have to see these two threats, uncontrolled AI and nuclear arms, this is a recipe for disaster. So therefore we thought that this was something very important for us to put forward and we're very proud that we achieved majority of the United Nations General Assembly voting in favor of this resolution.

Does that mean it will happen though, does it? I mean, you've got to get the right people on board to actually implement it.

Yes, of course.

So are you optimistic that you will get what you all agree on?

Well, I think it depends, of course, on those countries that did not support the resolution and the people that make those countries. accept that nuclear weapons cannot be handled by artificial intelligence. This is a principle. Nuclear weapons cannot be handled by artificial intelligence. And this is something that at least we have to agree while nuclear weapons still exist.

Yeah, so David's saying, this issue of human agency and making sure that AI is, it never becomes the master and that although it's got all this benefit and so on, but We've heard Enrique Oltrua talk about how there is uneven access in Mexico. And I remember Kofi Annan years ago talking about the internet, saying the digital haves and the digital have nots of the world and so on. So how do you approach this in Sierra Leone to make sure that you don't end up having those who are left behind?

I really like what you said about technocracy. Some concepts that we have rolled out or speak about in Sierra Leone is techquity, so using tech to bring about equity, and there are two things. One is we developed what we call country as AI lab. And the point is how do we use AI to deliver services for people to help government leaders, private sector, and people in the classrooms make better decisions to do the things that they are going to do, make humans better. If you deploy country as AI lab, a safe environment where private sector and government are working in this sandbox, then we get the best solutions when we think about those who are not in the room, radical inclusion. If you build for the poor, the marginalized, the poor girl, the pregnant girl, you're going to develop a better solution for everybody else. And this is real and it's in practice in Sierra Leone. I'd like to take a small step back into history. I've been mentioning in 1800s, extracted raw people from Africa. In 1900s, they extracted raw minerals from Africa, and I think in other parts of the world. And in the 2000s, early parts of the 2000s and now, they're extracting raw data, which is an ingredient for all the different industrial revolutions. And in this case, data is so important for AI. That's one. And if you hold the thought, and you-- World Economic Forum said that about the top 10 million websites in the world, 90% of them were in 10 languages. Africa has 1,500 to 2,000 languages. And we know that the diversity of language, in this case, is a powerful tool for AI models. So if you think about TechQuity, if you think about radical inclusion, if you think about creating value on the continent, the solutions that you build will be better for everybody else. As an example, using that country as AI lab framework, two examples in the classrooms in Sierra Leone with Fab AI and Google DeepMind, within eight weeks, we did scaffolded learning in classrooms. and students who are engaged and this is asking questions in mathematics in Sierra Leone and in Northern Italy, they had a 2.5 year learning gain within eight weeks. So now for kids who are left behind in math and kids who are not in private schools, these are normal kids compared to the control, you're able to ensure that all of them can learn. And these values and these power of what AI can do is really important. And final example, our team in Sierra Leone, high illiteracy, worked with the people in the school for the hearing impaired to build models to allow for sign language for kids who now in normal school cannot go beyond primary school. And so you are able to bring in a lot more of your people who are excluded when you think about the the the marginalized and radical inclusion.

Thank you so um Paula we we've heard Minister we've heard um the chief minister from sary and David sang talk about um data and data sovereignty how are you approaching this in Rwanda because there are there is a lot of discussion about having Regional data centers so I want to ask you that and also whether you have to take into account, particularly in Africa, the fact that data centers consume a lot of power, a lot of water. So there is that measure to the kind of future we choose in terms of the data centers, and that should be much more transparent in terms of, you know, the quid pro quo and the fact that they do consume so much water and power and so on. So if you could address the issue of data sovereignty and also having this lens on how we measure AI.

Before I address that, allow me to just pick up on the question that Minister David Sange responded to. When you think about the AI statistics around the 2.2 billion people that remain unconnected, majority of those are in parts of Africa, in our countries, and that is people who do not have any smart device, even as simple as a basic entry smart level smartphone. Their digital literacy skills are very limited, and in many ways, they are not covered by broadband connectivity investments that countries are making. Now, divide that we've been dealing with for many years and trying to close. Then comes the intelligence age where we're thinking about AI with its potential, granted, to accelerate development, to accelerate the closure of these gaps, but also with a risk that if we're not careful, it will exacerbate that divide. And to the haves and have nots that you alluded to earlier. So, as we think about deploying AI.

Let's put it out to you.

As we think about, you know, I may have to just hold it.

Yeah, do that.

As we think about implementing AI, what that means for countries like ours, if we build an AI solution for a farmer who doesn't have a smartphone and whose digital literacy skills are very limited, what's that going to look like? Because AI is not going to come and replace this farmer. So these are some of the design parameters that we're looking at from a human development perspective, but also being able to contextualize our specific context as African countries. Now that brings me to the question you asked around data, data sovereignty, and the whole approach to data centers and compute. One, Rwanda's approach from was to put in place our data protection and privacy law because we wanted to make one we're giving agency to the citizens over their data they decide how to use it who uses it and all of that but also the era where we're using lots of data to create valuable solutions to our citizens I think the key question then becomes because when we think about sovereignty many people have interpreted it as I host all my data in country but you may not have that capability and so the question should be can my data move with my rules? This is where the concept of digital embassies over the past years has become prominent. Even for countries that don't have data center infrastructure, they're still able to host in countries that have similar rules. And the last part that I wanted to come to in the interest of time is how we're thinking about the regionalization of compute hubs. To your point, we have countries that are having to ask themselves, do we want have data centers when we haven't even finished with our electrification plans. Rwanda, for one, we have our electrification rate is at 86%, and that's why today we also have a nuclear power program in place that can power some of these data ambitions and AI ambitions that we have.

Thank you. Final question to all of you, and to you first, Enrique, is, this is AI and the future we choose. How do you interpret that question? Briefly, if you would, thank you.

I think what I would say is basically it is about ethical values. It is about how do we include more people in and kind of bridge that digital gap that we were not able to bridge during the previous technological revolution. I think AI needs to bring people together while respecting all the cultural diversities of our people.

Thank you. David Sange. AI in the future we choose.

When we say humanity, the question is, who is humanity? When we say values, the question is, whose values? When we say ethics and morals, whose ethics and morals? I think, obviously, humanity, we can define it how we want. But as we build this future for AI, we focus on the most marginalized, the most excluded, the most unseen, the most vulnerable, the solutions and the answers we get to for all humanity will be better.

Let me help you. We're in the United Nations, there's the Universal Declaration of Human Rights, and we talk about which values.

But on that, we are still in the same UN and the point about humans and who gets to say certain things is not quite clear, universal.

That's something else. All right, Paula Ingaberry, final word quickly, please. Thank you.

Final word. AI and the future, for me, it's about partnerships that are going to help transfer and build capabilities for countries, partnerships that are going to be critical for how we preserve our future choices and also deliver what is intended.

Thank you very much indeed. Thank you very much indeed to this panel. Thank you. And it's so tight, everything. Thank you. Thank you so much. Thank you. So if I could invite for our final panel, Amir Siraj, astrophysicist and concept pianist, and also Joseph Gordon-Levitt, actor, American actor, to come to the stage when we're going to be discussing creativity and humanity in the age of AI. Thank you. if you come up gentlemen hello I'm.

Very nice to meet.

You to meet you good hi hello hi

so you know creativity is really goes to the heart of what makes us human doesn't it because although we know there are lots of you know, lots of songs and voices we can hear which are generated by AI. I know I love opera. I wouldn't like to go and see an opera where you've got a load of robots singing, even if they get the best pitch and so on. So This goes to the heart of creativity, doesn't it? We admire human endeavor. We admire the fine actors, the pianists, because I know you're very good. And we admire the dancers and so on. I know everybody enjoyed watching those robots running in China, but you'd like to see the real thing. Usain Bolt, that kind of thing. So human creativity endeavor, that's what we really admire. Do you think we'll ever lose that? Because, I mean, we've seen Tilly Norwood, this AI star. I mean, we'd rather have a real star like you.

Well, I was going to let you go first, but then she brought up Tilly Norwood. So you got to address it. If you haven't heard, Tilly Norwood is this piece of software that's being branded as an actress. And look, I think your question is a really important one. Can AI be creative? I think the first thing to remember is that AI is not a person. I've heard it referred to as a new species. I think that's an illusion. AI, as we know it, is an algorithmic recombination of human inputs. It's taken everything that everyone's written, all the films, all the songs, everything on the internet, and when I say taken, I mean, you know, taken without consent, without compensation. And then a very sophisticated and expensive to run algorithm has recombined all those inputs to generate these new outputs. So an output that's branded Tilly Norwood is not an actress. In order for someone to give a performance, there has to be a performer. When you think of a character like Donald Duck, for example, you don't say that Donald Duck is an actor. You don't say that Donald Duck gave a performance. You say that Donald Duck is an animated character that was created by a number of different artists, writers, animators, voice actors, et cetera. And all of those artists are given credit for creating Donald Duck. The product known as Tilly Norwood is no different.

If you watch her on the film, she's going to be starring in a film soon, isn't she? She looks pretty human. Most people will be seeing her on a screen. I appreciate people may not want to see her in the theater, but that is a bit of a threat. But anyway, let me come to Amir Siraj and how you interpret this issue of creativity in the age of AI.

Yeah, well, there is one thing that concerns me quite a bit, which is, you know, I have spent over 20 years of my life learning to play the piano and honing my craft, right? There were no shortcuts I could take. There was no way around this. And I'm really grateful for that because those years built me into the person who I am. And the music is almost a byproduct. And so, you know, imagine all of the things that we're now seeing being automated, right? Reading, writing, reasoning, making an argument. And that list is growing day by day. And what concerns me is the possibility that children can be handed the result at the beginning and never have to go through the years that builds a person. Now, I am somewhat less concerned about AI's impact on music for a very specific reason. You know, for example, I was last year at UNGA about three years ago playing with Yo-Yo Ma at the Gold's Lounge for the SDG halftime. And there's a recording of that evening, and I've listened back to it. Now, the recording is accurate, but there's something missing. You know, whatever made that evening matter to the people in the room is absent, because that's what a recording is. And so, you know, every one of these systems is trained on records, right? A book is a record of a thought. A score is a record of a piece of music, and a record can be entirely accurate and still leave out something that's specific and critically important. And so that's what I think perhaps is the clearest description of what these convenings were all about. Because, you know, in every city, the things that people said that mattered most to them and that they were most afraid of losing were things that only exist when people are together in community or that require years of learning. These engender values such as judgment, trust, imagination, courage. And these won't magically appear in AI systems, and it's not easy to engineer their protection either. In fact, the place that they need to be protected is in how we choose to live alongside these systems. And that is a decision about values, and it's being made now with or without us. And to channel the DSG here, I certainly believe that those kinds of decisions need to be made deliberately by we, the peoples of the world, with the broadest we possible.

That's your astrophysicist hat there as well, but just sticking with this issue of creativity, You know, you could say, I know you're obviously a marvelous classical musician, but you could say compose a new symphony in the style of Beethoven and AI would do that. And I could sit there and enjoy it. I may not admire the person, the machine that's done it, but it doesn't stop me enjoying it. If there's a beautiful piece of art, you could still enjoy it, couldn't you? But it's this issue of admiration. So, I mean, do you agree with that, that we could admire great art or great music composed or created by AI?

I think a couple of things happen when you enjoy a piece of art. One, I think you can enjoy just the art itself. But I do think something else often happens, and that's that you feel a connection to the person who created the art. And maybe there's a difference between creativity versus expression. an AI might be able to, like you say, create a symphony, but it can't express anything because all it is is an algorithmic recombination of its inputs.

But I would challenge you there because, I mean, you know, I don't know, da Vinci dead for hundreds of years. Somebody draws something in the style of da Vinci. I mean, You could still enjoy it, couldn't you? But it's that issue of admiration that you don't have. There isn't the aura, rather, of the author, the aura of the composer, the writer. And that's very, very important to retain that aura. So it's like, you know, when you meet somebody who's a great artist, a great scholar, very learned, they have an aura about them, don't they? Which I suppose AI wouldn't.

They have a life, they have a perspective, they have a self. And a lot of these AI products are putting forth an illusion that they have that. In the very language that these products use, they refer to themselves in the first person. They say, I. products are named with human names. They say things like, how can I help you today? They say things like, I'll make you a list of this or that, or I made a document for it. There is no I. The word I means something. It means a self. It means a subjective-- it's literally the subject of the sentence, the word I-- a subjective experience. And this, I think, is technically not there.

So Joe's brought up this issue of collaboration, cooperation, communication, because that's often what inspires great art, isn't it? You know, you bounce ideas off in your industry. It's very collaborative, isn't it? You bounce ideas, you get inspired by people. I mean, Do you think that we have to nurture a different way of thinking how we encourage people to be creative by saying, don't just sit in front of your AI machines and do this, but go out, experience museums, art galleries, travel abroad, see different things so that our knowledge, our creativity is much more based on experience. so that you see the passion dancing in people's eyes, which you can't see with AI.

I think that's absolutely right. And I think that harkens back to a lot of the conversations that we've been having in these convenings. It was interesting. One of the first things in Cairo that everyone agreed upon was this question that sort of became an unofficial motto of the convenings, which was, If we can't stop this, can we steer it? And the second area of agreement was that no one was happy with who was steering. These AI systems are, by default, being developed and shaped by a few companies. And nobody at any of the convenings was happy with that situation. And now, I do want to put this in context because much of what's been in the news in the past couple of weeks has been about safety, has been about can we control these systems. And while that is critically important, it is a prerequisite to the kinds of conversations had during these convenings. Because even if a system can be controlled, it still needs to be pointed somewhere. referring to this kind of intentionality that you're talking about. We still need to decide how we want to exist alongside these systems. And so, you know, in Cairo, a group of us over lunch started to try to think about how can we distill some of these discussions into guidance. And, you know, it became as 10 lines in this very tiny notebook. And by the time Rome was over, it had developed into this draft called Principles for Human Centered Artificial Intelligence. I believe it's on 50, page 51 of the report, which tries to lay out, you know, a draft of what a voluntary standard could look like for companies and countries. And in there, there's some possible red lines as well as some values. And the red lines came together quite quickly and changed very little from continent to continent. Things like a machine must never decide alone to take a human life, never be used against a child, never be making a decision about someone's livelihood or freedom without a person to appeal to.

Just staying within the realms of creativity, I understand that. Thank you for giving us a reprise of what was discussed there finally and briefly. Do you think in creativity in the age of AI, we may definitely see AI created art and an enhanced role for, I mean, AI already exists left, right, and center in movies, doesn't it? I mean, you know, it's everywhere, isn't it? But, you know, will we see AI generated art? But there will also be human created art and we will value that even more because there'll be all the scripts and whatever rubbish written by AI.

Here's where my mind goes when you ask me that question. And it might sound boring, but I think it's actually deeply important, and that's the economics. How is the money working? That historically, there has always been an economic incentive for human creativity, human ingenuity. And I'm not just talking about in the arts. I mean, any human endeavor that's creative, whether it's a work of journalism or a work of science or a work of engineering or any human creativity has had an economic incentive. right now, what's happening is a few enormous businesses are taking the work of everybody, and not just artists, but everybody, taking that work, using that work to train their AI models. In other words, algorithmically recombine all that work to generate these outputs, and claiming that all the people whose work was used to train those models deserve 0% of the economic upset.

Was that strike, wasn't there a few years ago in Hollywood?

There are a litany of lawsuits. There's a lot of debate about this, yes. And because I think this basic principle is deeply important to the future of how everything works. There's actually a right in the UN Declaration of Human Rights. It's called the right of remuneration. It means that if you do some work, you deserve to get paid for it. We are on the precipice of entering a new kind of technological revolution where that right is fundamentally violated, where everybody's work can be-- The authorship is devalued. Yes, and everybody's work could be just vacuumed up into these AI models. And so what does that lead to when you take it a few, even just a few years into the future? What that means is that all the money and all the power concentrates into the hands of just a few big businesses. That is not a free society. That is totalitarianism. And the right of remuneration needs to be upheld in order to avoid that future.

And will it undermine creativity, human creativity? You think it will. Or will it just give it greater value because of its scarcity.

I think if there is no economic incentive for human creativity, it is under threat of not surviving.

Amir.

Yeah, the economics are crucially important and so is what makes art art. I mean, it's an open question and we're having to debate it right here and in many, many places across the world. And, you know, I think this merits thought about what is just about aesthetics versus what makes something a work of art. And I'll talk about this a little bit later about the Golden Record, but, you know, there's something very unique about a human being bringing their life to their work. And that's something that cannot be replicated no matter how beautiful the aesthetics are.

I know you perform with Yo-Yo Ma, the famous French-American cellist, right here in the UN. I mean, just very quickly, are you frightened by this kind of conversation and what Joseph just said?

You know, in some ways I am, and in some ways I'm not. I'm worried about what will happen to us as humans, chiefly, if we do not make decisions about the values. So, in each convening, you know, different kinds of values were emphasized. So, for example, in Cairo, the concern was brought up that we might be reduced to our data, right? In Mexico City, people talked about the resources that these systems consume. In Xi'an, it was about a child's imagination and the ability to learn things the hard way. And in Rome, in fact, it was emphasized the people whose work these systems are based upon, and if they can continue to do that work. And so, you know, and the last thing I'll mention here is that it's also very important to figure out how we verify that companies and countries that their conduct actually follows any stated or agreed values. And that, you know, we put forth an idea that perhaps the burden of proof should be on the companies and countries to publish evidence openly that third parties, the public, can actually check.

Thank you for giving us that overview. But just sticking with the issue of AI and creativity, we are the only species who make art for art's sake. That's one of the things that defines us. as human and long may that continue and long may you continue to act and not be replaced by some silicon substitute and indeed that you retain control over your image because that's something that people like George Clooney have said he's worried that in 20 years after he dies that his image will be used for promoting and doing all sorts of things but I think we all admire human endeavor as represented by the two of you thank you very much indeed thank you thank you Thank you. So that's all from me. I'm going to wrap, just bring Anne-Marie Hou, who is executive director of the UN Office for Partnerships, is going to just say some closing remarks and also tell you what's next. But I'll just end with one thought, which is a great heroine of mine who's no longer with us, Wangari Maathai, the great Kenyan environmentalist, who, not say this within the context of AI, but I think it could be applied. we're discussing all this, which is, you cannot enslave a mind that knows itself, that understands itself, that values itself. And I hope that that Wangari Maathai's words still ring and inform this whole debate about AI and human development. Thank you for your attention. Goodbye. Thank you.

Excellencies, colleagues, friends, and everyone joining us online from around the world, before anything else, please join me in thanking Zainab Badawi for guiding today's conversation with such curiosity and humanity. Zainab, thank you so much. I'm going to start with just a quick question since we still have our cards, and I only brought up one, so you know my answer. Did all of you learn something today? Raise your green card or red card. I mean, I think it was unanimous. There was so much depth to the conversation today, and that's what we need to be doing all around the world. I want to express my deep appreciation to the Deputy Secretary General, Amina Mohammed, for her vision and leadership in convening this global conversation, and His Eminence Cardinal Pietro Parolin, Secretary of State of the Holy See, for his inspiring statement in support and engagement. And thank you to everyone here and online for imagining that this conversation can take place and imagining having this conversation with your peers. Over the past several months, we've been to many continents, cultures, and disciplines, from Cairo to Mexico City, Xi'an, Rome, and now New York. After all the discussions about models, algorithms, inclusion, governance, technology, we keep returning to one question. What kind of future do we want to create together? And we keep coming back to three things, people, relationships, and hope. So today is not the destination, it's just the beginning. As we look toward the SDG Summit, which is gonna take place right here, same time, next year, we're working to ensure that these conversations will travel into all the communities around the world. with UN Resident Coordinators and country teams working alongside governments, academia, civil society, private sector, faith communities, and most importantly, young people. Similar to our convenings report, we want to capture all the voices like we did in the dispatches. And in every place, we gave a little animal to put on the table. It was the proverbial elephant in the room. We don't want or need any more elephants. So in Cairo, it was a camel. In Rome, it was a dove. Wherever community you're in, don't leave the elephant in the room. Make sure you're in a safe space that we can start talking about these very big issues that can be scary, that are so important. Because the future should not be shaped by a handful of people and a handful of rooms. This morning, Victoria opened our gallery with a story about possibility. As we close, we leave you with another story, this one about imagination, about a species that looked up at the stars and chose to send a message out into space. So please join me in welcoming back Amir Suraj, who's gonna share a story about the Voyager Golden Record, one of which hangs actually upstairs in the United Nation. And I'm really excited to welcome back Amir.

Thank you, Annemarie. I study the outer solar system, the part of our cosmic map that is still mostly blank. Out there right now is Voyager 1, launched in 1977, the farthest thing human beings have ever sent. Bolted to its side is a phonograph record called the Golden Record. You may have heard of it. It was built to last a billion years, and it carries 90 minutes of music and greetings in 55 languages. It's a self-portrait of humanity made for whoever might someday find it. The team that made it was led by Carl Sagan, together with Ann Druyan. Sagan was my PhD advisor's PhD advisor, so this record sits very close to home for me. One part of the record that took a very long time to settle upon was a piece of music from China. Near the deadline, a musicologist told Andrian that the piece should be flowing water played on the guqin, a melody some 2,500 years old. She called Carl with the news and left a message. He called back an hour later, and by the very end of that call, they were engaged. Carl passed away in 1996. Anne is fortunately still with us, and I know her, and she's very excited about the legacy of the Golden Record and how it is shaping our discussion around AI today. Now the melody of flowing water carries a story of its own. Bu Ya was a master of the qin, and Zong Zhiqi was a woodcutter who would listen. When Bu Ya played with mountains in his mind, Zhiqi heard mountains. And when he played thinking of water, Zhiqi heard water. When Zhiqi died, Bu Ya cut the strings of his instrument and never played again. The Chinese language kept a word from that story, zhiyin, the one who knows the sound. And today, it still means a true friend. I keep returning to that word. AI systems can already produce the sound in any style, in any voice. But Zhu Qi brought a whole life to the listening. When that was gone, all of the music in the world couldn't replace it. Now, if the melody is about listening, the golden record itself was about choosing. A handful of people had less than six weeks and a single chance, and they weighed every minute of sound because whatever went aboard would speak for everyone. AI systems are becoming the next version of that record. But whereas the golden record was deliberately constructed, These systems are being assembled from whatever happened to be written down, digitized, and scraped. And the deeper choices, what they are optimized for, and what they are permitted to do in our names, are being settled the same way, by whoever happens to build them. The Golden Record was addressed to strangers we will never meet. These systems are addressed to our own children. And they will do far more than carry a portrait of us. They will teach, they will advise, and they will act on our behalves. That is our responsibility. And the report in front of you is where we have tried to start. Now, when aboard Voyager was a recording, and this morning we're very lucky, it will be played live right here by Wenrui Shi. So let me leave you with a thought. If AI were our generation's golden record, what would you put on it to represent humanity?

Thank you.

Oh.

Thank you so much. That was incredible. Thank you, everyone, for being with us for AI and human development. Let's start that conversation now and have that with your families, with your communities. And again, keep your imagination and think about what it means to be human in the age of AI. Thank you.