SECRETARY BLINKEN: Thank you. Thank you very much. It is wonderful to be back here at the New York Public Library. And I have to say at the outset that I very much share my good friend and colleague James Cleverly’s sense of optimism. It is an occupational hazard. In fact, I think we’re both reminded – almost daily – about what the late John McCain used to say: It is always darkest before it goes completely black. (Laughter.)
But here at this library, in fact, you consistently for so many years, so many decades, are actually bringing light to the greatest challenges of our time, and light through knowledge onto what we need to carry our societies forward. It’s a living demonstration here at the library of how information and ideas, when freely accessed and exchanged, can actually empower people and empower communities.
And that’s exactly what we’re focused on today: How can we use artificial intelligence – the technology that is revolutionizing the production of knowledge – to actually improve lives; to make the world a little bit safer, a little bit healthier, a little bit more prosperous; to help eliminate poverty and hunger; to promote health and access to quality education; to advance gender equity; to save our planet.
So let me start by thanking our colleagues from Japan, from Kenya, from Morocco, from Singapore, from Spain, and from the UK for bringing this event together with us. To our Moroccan friends, especially in this moment, we send our deepest condolences for the loss of life and destruction caused by the recent earthquake.
I’m grateful, also, to the many partners we have here from civil society and the private sector. If we’re going to succeed in directing AI toward achieving the Sustainable Development Goals, we need – we need the full and robust participation of technologists who are developing these systems; of governments and organizations that are, as you heard James discuss, shaping their norms; and the citizens, the communities that are affected by their growing applications.
So today, as everyone in this room knows, we are in an entirely new era of AI. Powerful generative systems like GPT-4 – orders of magnitude more capable than their predecessors – are vastly increasing the capacity of machines not just to process information, but to create original content, to perform complicated tasks, to solve the most vexing problems.
We all know as well that this new era brings potential harms and hazards, including AI systems producing false information, reinforcing bias and discrimination, deepening inequities, or being misused for repressive, destabilizing, dangerous purposes.
In coming days, months, and years, it will not be our technical capabilities that define our future, but how we choose to use them. Our responsibility is, of course, to minimize the risks while maximizing the potential for AI to tackle some of the world’s biggest challenges.
And that has to include achieving the Sustainable Development Goals: the goals that we, as an international community that’s come together as it does every year this week, goals that we set for ourselves in 2015 to bring about a better future for all our people; and as we said, a “shared blueprint for peace and prosperity for people and the planet, now and into the future.”
But here’s the reality: We are falling short. We’re halfway to 2030 – the deadline that we gave ourselves for hitting these targets – and yet we are on track to achieve just 12 percent of the SDGs – 12 percent. Our progress seems to have plateaued on half the goals that we set. On nearly a third, we’re actually moving backward; we’re regressing.
That should simply be unacceptable. And it is. And that’s why we have to use every tool at our disposal – including artificial intelligence – to get the SDGs back on track.
Now, experts estimate that AI could advance progress on nearly 80 percent of the SDGs and their targets. Already, we’ve got AI systems being deployed to forecast extreme weather events and the impacts that they have; to improve agricultural productivity to fight global hunger; to predict, to prepare for, and respond to outbreaks of disease and new viruses; to build the clean energy infrastructure for a healthier future.
So we’ll hear from – directly from colleagues about their own efforts, some of which are on display in the exhibits around this room, and I really commend them to you. I got a brief preview, and it’s really worth taking a look.
The United States is committed to supporting and accelerating efforts like these, to fostering an environment where affirmative AI innovation can continue to flourish.
That’s why we’re helping to lead the way forward on AI governance. Just over the past year, we’ve built a Blueprint for an AI Bill of Rights to guide how automated systems are actually used as well as designed. We published an AI Risk Management Framework to improve AI user protections. We issued a set of voluntary commitments, developed with the top AI companies – including those with us today – to increase safety, to increase security, to increase trust for advanced AI systems. More leading companies are getting on board: seven additional companies joined this framework, these voluntary commitments, just this past week.
Now the work that we’re undertaking is to work with partners to translate these commitments into an international framework to govern AI, including through the Japan-led Hiroshima Process in the G7, the UK’s AI Safety Summit in November, and India’s leadership as chair of the Global Partnership for AI.
We also know this: If these norms are going to be effective, we have to bring a wide range of voices and views to the table, bring them into the discussion, including developing countries. And we’re committed to doing just that.
As we work to manage risks, we can and we must also maximize the use of AI for the greater good.
In the voluntary commitments, companies pledged to create AI systems focused on addressing society’s greatest challenges. We’ll work with companies to help them make good on that promise, as we strengthen the partnership between developers and governments more broadly on AI.
At the same time, we can deepen our cooperation across governments to foster and also incentivize more affirmative uses for artificial intelligence. Right here at the UN, we’ve helped launch the AI for Climate Action and are strong supporters of the International Telecommunication Union’s AI for Good initiative.
You’ve got so many parts of the United Nations system that can bring important perspectives to bear on this effort. We’ll hear later today from the leaders of the International Telecommunication Union, the World Health Organization, the United Nations Population Fund, the UN secretary-general’s tech envoy. And I want to thank each of them for being with us today, but also for the work that you’re doing every day to advance an affirmative AI future.
Today I’m also pleased to announce a new $15 million commitment by the United States to helping more governments leverage the power of AI to drive global good, focused specifically on the Sustainable Development Goals.
Gatherings like this one, where we bring together and learn from an incredible array of ideas, expertise, backgrounds – it’s a critical step to getting this moment right. As President Biden likes to say, we are, in so many ways, at an inflection point, where the decisions that we make today will shape the world not just for years to come but for decades to come. And AI is probably the best, most compelling example of that.
The decisions that all of us make collectively today about how we shape the future of AI – that is going to define our world for decades. This is an immense responsibility, and it’s a responsibility that we have to take together as governments, the private sector, civil society, individual citizens, and communities.
If we don’t get it right, I think we have a pretty good idea of what some of the risks are. But if we do get it right – and we can, and we will, and we must – the opportunities are almost endless.
Thanks very much. (Applause.)