Bill Gates: Technology is a world in need of governance

By Roberto Cetera

“We are not paying enough attention to the risks” of artificial intelligence “and we’re not doing enough to minimize the harms. We are also not doing enough to take advantage of the current opportunities with AI,” says Bill Gates, cofounder and former CEO of Microsoft, in an interview Vatican News.

The tech leader also spoke about Pope Leo XIV’s encyclical Magnifica humanitas, saying “What struck me is that the Pope isn’t saying ‘stop AI’. He’s making a more careful argument: that we need to actively preserve what’s human in the face of it. That lines up with something I’ve been arguing myself”.

Read the full interview (lightly edited for length and clarity):

Q: Mr Gates, in recent weeks, many voices from within the technology industry have raised concerns about the potential dangers associated with the further development of artificial intelligence. What is your view?

Bill Gates: It is striking how people inside the industry are voicing their concerns. These are people who are competing with each other and disagree on many issues, but here, they’re saying the same thing: This is not something the industry can self-regulate, it must be led by neutral actors, primarily governments. The trouble is most people in government don’t fully understand how the technology works.

It’s my view that we are not paying enough attention to the risks, and we’re not doing enough to minimize the harms. We are also not doing enough to take advantage of the current opportunities with AI. The [Bill and Melinda Gates] Foundation released our annual Goalkeepers Report last month, and it made the case that AI could be a great equalizer or widen existing gaps between rich and poor.

AI can do things like let rural doctors use a smartphone as an ultrasound machine, provide personalized tutoring that adapts to how a student actually learns, and offer advice tailored to a farmer’s own soil and weather. But none of that happens automatically.

Left to the market, the most capable tools will get built first for whoever can pay the most, not for whoever needs them most. That’s why we committed a billion dollars over the next two years to push AI toward health, education, and agriculture in the places market forces would reach last. The risk and the opportunity are moving at the same speed, and we have to act on both at once.

Q: More generally, how do you see the future of an increasingly technology-driven society?

AI is different from any other technology we’ve seen before. With past innovations like the PC or the internet, the pattern has started with some major disruption, but eventually it works out, and people end up with better tools. AI is different. It’s more like an evolutionary event. We’ve created a kind of intelligence that in many ways matches or exceeds human cognition. Treating it as just ‘the next PC’ causes people to underestimate it entirely.

We can think about it as two different periods. Right now, we’re in a transition where we don’t yet have the upside at a large scale everywhere, but we do have real disruption. That’s truer in rich countries than in lower-income ones.

Advertisement

In the places where the Foundation conducts our work, the benefits can arrive faster than the disruption does, because there was so much unmet need to begin with. That only works, though, if we fix things like how most AI models were trained overwhelmingly on English and badly underserve the billions of people who speak other languages.

So real change is required during this period, on multiple fronts, including how we tax labour versus automation, which jobs we deliberately reserve for humans, and how we make sure the technology actually reaches the people who’d benefit most, not just the people who can pay the most.

Then, on the other side of that transition, if we get the guardrails and the access right, we could see a world of genuine abundance: AI and humanoid robots handling food, shelter, and other basics.

That period will raise a totally different set of questions about meaning and how people spend their time. My focus right now is getting us through the next several years without a catastrophic bioterrorism attack or cyber-security event, and without leaving huge parts of the job market, or huge parts of the world, behind.

Q: President Trump, by contrast, has minimized the potential risks of further AI development. Do you agree with his position? What role should political leaders play in this debate?

This is technology that’s more dangerous than nuclear weapons, and has the potential to be more beneficial than nuclear energy. Adding basic safety checks costs us almost no speed, and it cuts the risk a lot. That’s a trade worth making.

Government leaders need to first agree on the basic facts, that there are real dangers here, and real benefits, and then discuss details of the response. We need more AI expertise inside governments, governments working with industry, and governments working together, because no country wants bioterrorism or a cyberattack that takes down its own infrastructure.

I worry that it may take a serious cyberattack or a bad bioterrorism incident before the political system takes this as seriously as it needs to.

Q: Is there any connection between the risks associated with the development of AI and your decision to step away from any operational role at Microsoft?

No, my decision was a gradual, planned shift toward philanthropy and my work at the Gates Foundation, which happened more than a decade ago.

Q: Have you read Pope Leo’s encyclical Magnifica humanitas? Do you agree with its reflections on artificial intelligence?

Yes, I read it, and I wrote about it in my own memo in August. I was fascinated by it. It lays a strong foundation for the work that needs to be done. What struck me is that the Pope isn’t saying ‘stop AI.’ He’s making a more careful argument: that we need to actively preserve what’s human in the face of it. That lines up with something I’ve been arguing myself.

There are certain roles, like caretaking of children, serving as a judge or a juror, and some parts of medicine, where the relationship itself is the point. We shouldn’t hand those over just because a model becomes capable enough to do them.

I am very glad to see the Catholic Church, and other religious groups, so engaged in the AI debate; I believe their engagement is critical to ensuring that AI serves everyone in the most equitable way.