Washington and Beijing have just opened talks on super intelligence. The leaders of the top AI labs are asking to be regulated. And Bill Gates, technology's great optimist, warns that AI could drive events causing a billion deaths. Today, Africa.com has an exclusive interview with him to discuss his concerns and aspirations for AI and what it means for the world.
Thank you very much, Bill, for doing this interview. If we could start by talking about the essay that you've recently written that really speaks to the world and speaks to themes that are relevant for the entire globe, Western nations, Africa, developing nations alike. And one of the comments that you've made in your essay is that these are very turbulent times for AI as it is embraced and that we have to make important decisions now because AI could either be a source for good. I know that you are an enthusiastic AI user, or it could also be the greatest source of injustice. Can you talk about that from a global perspective?
Yeah, I'm concerned that people's awareness of how AI is a very different technology than what's ever come before. You know, its power to help improve education, to improve health, is much greater than I think people recognize. But also, as it works through various parts of the job market, and as people are using it, you know, kind of as a having relationships with the AI, you know, I think some of those downsides are not recognized. So my memo really was a, you know, a loud cry to say that it shouldn't just be the industry talking about those challenges. We've got to get all of academia. We've got to get think tanks. We have to get, you know, citizens, and then eventually, of course, the politicians, because many of the ways that we minimize the negatives and accelerate the good things work through politics. And you know, I just wasn't seeing the kind of response that is necessary here on either the good or the bad. So, you know, I can't overstate what a profound technology AI is, even compared to things like the PC and the internet that I had a chance to work on.
So, Bill, you stand at a unique nexus. You have tremendous credibility on this topic as one of the most significant tech players in history, but you also are one of the greatest philanthropists in history. And so your ability to make this call to action really is quite profound and should be heard by both sides of this equation. The question I have for you is, as you think about academics, think tanks, and people in the lesser developed part of the world, much of what is happening right now is happening at such a fast pace and driven by the tech billionaires in Silicon Valley. How can you get these other players to insert themselves in conversations where they're generally not invited?
Well, I'd say the universities are partly where you're going to have a lot of advanced users and a lot of deep thinking. And you know, society and governments engaging with young people at universities who are doing cutting-edge work will be partly how we get the broad engagement that's missing today. The, you know, the models themselves are kind of neutral. And you know, I do think the industry has talked about some of the dangers. I'm not indicting the industry. I'm saying that it's outside the industry where the lack of engagement is going to mean we're not ready, and therefore we won't do the positive acceleration or the minimizing of the bad that it's really called for here.
Well, I know through the Gates Foundation, you've been actively involved in a number of really great use cases for AI on the continent, in healthcare, and in education. Are there some that in particular you'd like to focus on?
Well, the one that's actually going the fastest is agriculture, because understanding which seed to plant and when, you know, getting good weather data, getting good price data, you know, taking a photo to see what crop disease that you're dealing with. The lack of information in Africa is pretty dramatic, and yet it's a meaningful part of the economy. And raising productivity there will be a key to economic development. You know, that you have more output, that you free up labor for other activities, you know, as was seen in the past in many Asian countries. So that's one that, you know, we have more users today than in health and education. You know, partly in health and education, you know, making sure it's done just right, making sure you're not making mistakes, is the stakes are a bit higher than in agriculture. And yet, you know, we need to get the data into the systems in order that those also move along. So, you know, health is what I was looking at when I was in Rwanda. You know, most of it was saying, okay, for the patients, what does this experience look like of connecting up to the AI? How good is the African language support? And then what's the integration into the system where Rwanda's pretty strong about making sure they're really gathering data and therefore seeing, you know, where are the deaths, where is the supply chain working, where are the workers not well trained. You know, that commitment to have good data to feed into the AI is a key enabling thing that may hold us back in parts of Africa.
In your essay, you said one of the things that we have to get right is that these tools have to be grounded in local data and local languages. And so coming to that point, and then let's talk about just sort of the realistic return on investment to the extent that you have, you know, small countries, small populations of various that speak a particular language. How do the incentives get put in in order to get the investment for that local data and those local languages if it just isn't the same as it is, as you know, preparing these and training them on US data in English, for example.
Well, definitely there's a market failure there. You know, the profit motive won't drive full super high quality coverage of African languages in these domains. So, you know, we're putting together a group of companies. You know, when we're in New York for the General Assembly week, you know, we'll be more concrete about how the pieces are coming together. But you know, the goal is pretty straightforward. It's to be able to take any of the African languages that you say have a million speakers and have the quality, as you're talking about health or education, of that interaction be very close to how good it is with English. It's achievable. We have to orchestrate hundreds of millions of dollars, and we're in the process of really pulling that together so that, you know, we can show that over the next two years that that barrier has been taken care of.
Let's just talk about jobs for a moment. This is of course an important topic in every economy, especially so as Africa as it looks to integrate into the global economy going forward. And so you've spoken a lot about how there are many jobs that are being taken away by AI. How should Africa be thinking about this as an opportunity perhaps as opposed to a threat?
