Call the meeting to order.
So we have growth at around 3%. It used to be 3.8%. And here is the good news: AI can lift growth by 0.8%.
Who benefits is not determined only by whether you were born in a more developed country or not. It is determined by the kind of job you hold.
So, hello everyone. I'm David Ignatius, a foreign affairs columnist at the Washington Post. Thank you all for joining us on what is a very busy day, even by New York standards. I'm pleased to be joined today by the IMF's managing director, Kristalina Georgieva, who's led the IMF since 2019 and who began her second term last year. I've had the pleasure of talking with Kristalina often over the years, and I'm especially pleased to have her here. Kristalina, let's start with the obvious question for our session today, which is the fundamental impact of AI on the global economy. You at the IMF issued what I would call a groundbreaking report last year in which you predicted that AI would affect almost 40% of the jobs in the world and 60% of the jobs in the advanced world. I want to ask you, a year later, how that prediction looks and what you're beginning to sense about the world's ability to adjust to those enormous changes.
Thank you for having me here. Last time we were on a screen, the last time before it on a screen, much nicer to be in person.
We have been analyzing both the enormous advantage of artificial intelligence for an economy that has been growing very slowly, as well as the risks that are being created on the side of the advantages. We have been underwhelming in terms of growth. On one level, I would say the world economy growing at 3% is not so bad; it has proven to be very resilient. But historically, the two decades before the pandemic, growth was on average 3.7%. So what can artificial intelligence do? It can give us back this missing 1%. Our analysis shows that there is a very high probability that increase in productivity, which is what we have been lacking — the fact that the economy is so sluggish is because productivity has not been growing — that can be compensated by artificial intelligence. And the early signs are that indeed there are productivity gains. I see it in my own organization that are certainly very encouraging for a world that needs to grow faster if we are to meet the aspirations of people.
You mentioned the 60% in advanced economies, 40% on average. It actually goes down to 26% in low-income countries. Jobs that are impacted by artificial intelligence are either enhanced, or diminished, or eliminated altogether. What we are seeing is that this is happening, and it is likely to be happening faster with investments in artificial intelligence, especially in advanced economies. So this is like a tsunami hitting the labor market, and the question is, are we ready for it? You mentioned the report we published. It has a very interesting component. We decided to rank countries on how well they're prepared for the tsunami, and we came up with a very interesting result. We looked at digital infrastructure, labor markets, skills, innovation and how it transmits in the economy, and regulation and ethics. We have on top only advanced economies, we have at the bottom only low-income countries, mostly conflict-affected countries, and the distance between the best and the laggards is an accordion opened very widely.
So you can decide to be an optimist and say, 'Oh well, low-income countries, they're not going to be hit by the tsunami so much.' But then you think about the productivity gain — what is going to happen to these economies in comparative terms? So the story of artificial intelligence, as it shapes up now, is one of three pieces. First, overwhelming dominance of financial flows in the United States. Last year, 75% of venture capital in AI was invested here. This year, I'm told it is now up to 86%. So there is this very, very powerful sucking sound of capital going into the United States, creating a lot of excitement. I'll come back later on a risk that may be coming. Second, it is a story in which the relative preparedness seems to be higher on the physical infrastructure side and lower on the ethics and regulation side.
Let me just break in to ask about one implication.
Just a very interesting story, I cannot sit. I hope everybody in the audience remembers the AI tsunami description because that's a powerful one. But the IMF worries every day, and has since its beginning, about the gap between the better-off countries and the worse-off countries. And as you describe the world we're heading into, I worry about a much more radical break between the advanced and less advanced countries. And I worry about AI deserts in the world. You know, this tsunami — changing, empowering, hopefully enriching, but maybe missing parts of the world. Speak about that, the problem of the divide.
The risk is real. When you think about the fundamentals for artificial intelligence, you start from digital infrastructure and access to electricity. Half of the population in sub-Saharan Africa has no access to electricity. So if you want to get the power of your phone, well, you have to power the phone first. And that is a problem that my dear friend Ajay Banga at the World Bank is absolutely determined to focus on. They have a program to bring access to electricity to 300 million people in Africa. This is half of those that do not have access today. So the investments in basic infrastructure, and then of course investments in education and skills, in human capital — they are falling behind on everything, but they're also impacting the ability to make progress on artificial intelligence. Could that be accelerated? This progress be accelerated? Yes. But if you look at the speed with which advanced economies are moving in the world of artificial intelligence, I worry that the risk is more of the divide deepening.
We also have occasionally some positive signs when young people in developing countries come up with brilliant ideas and they solve problems. By the way, one thing that I have observed with great admiration is that in developing countries, entrepreneurs are much more likely to concentrate on using artificial intelligence to solve the problems of their communities — access to clean water, ability to transmit knowledge to remote communities. So there is that element, but on the broader spectrum, the risk is still more tilted towards the divide deepening. How can we help countries? For our advice to developing countries, to emerging markets, to low-income countries, is concentrate on these two key impediments for you: basic digital infrastructure, and please make it public digital infrastructure. If you make it private, you suffocate your chances to catch up. And concentrate on skills. One of the good things about artificial intelligence is that it actually accelerates knowledge transfer. Focus on that. Don't think first and foremost about profitability. Think about your ability to be part of this AI story.
So, one of the things that the IMF worries about is global financial stability. And with the enormous levels of investment that we're seeing in AI, this sort of desperate scramble among investors and companies to build the infrastructure — you know, this morning you read of the hundred-billion-dollar investment that is contemplated in OpenAI by Nvidia. You see estimates of $3 trillion over time in this sector, and you begin to wonder about — not about the AI boom, if you're a journalist, but about the AI bust, about the danger of financial instability caused by this enormous wave of investment. What do you think?
