For nearly half a century, Bill Gates has been the person who embodies our hopes for the power of technology. It was the personal computer and Microsoft's operating system that made Gates the world's richest person for several decades.
Now he says everything has changed. In recent years, on two occasions, he has seen AI develop in a "shocking" way, which scares him. He now wants to spend the rest of his life trying to tame the harmful effects of AI, in the hope that the development can still be turned to something positive.
- For the first time in my life, I no longer see innovation as something only positive, he says.
It's Tuesday morning on the American West Coast, and on the screen, Bill Gates looks exactly like Bill Gates. The dark blonde bangs, the steel-rimmed glasses, the white shirt under a dark cardigan. He always seems to be dressed as if he is ready for the chilly fog that usually sweeps in from the Pacific coast of Seattle, where he has lived all his life. He's sitting at home in his house overlooking Washington state's mountain peaks and spruce forests, surrounded by no fewer than three associates who seem to be there to make sure the interview doesn't go off the rails.
"About three years ago I saw Chat GPT for the first time and it shocked me that it was apparently very, very powerful"
It is Gates himself who initiated the conversation, under the pretext that he is so keen to warn of the risks of AI.
- About three years ago I saw Chat GPT for the first time and it shocked me that it was apparently very, very powerful. Last year I then saw the Claude models become extremely good at generating software code. Understanding software code is probably my top skill, given that I was completely focused on becoming a good coder as a teenager. I was totally obsessed with it. But now Claude is already better than me at coding!
Bill Gates was born in 1955 – ten years after the US dropped the first atomic bombs on Japan. The awareness that there is technology that can wipe out humanity in an instant has never been foreign to his generation. But he has also lived his life during a period when technology brought about radical improvements in our everyday lives and an incredible prosperity grew on the companies that controlled the products, not least Microsoft, which he himself founded in 1975.
These fortunes have brought with them political and cultural power. American technology companies are today richer and more powerful than many nations. They have also made it difficult to criticize the development of technology without facing resistance. If new technology is questioned, technology optimists often appear who remind us that innovation, in the long run, leads to more jobs, higher productivity and more prosperity for all. Any social and political side effects have been marginalized as acceptable costs of these riches.
But with AI, something has changed. In recent years, the loudest warnings about development have not come from dogmatic skeptics or retrogrades. They come from within the industry, often from the most experienced voices – researchers such as Geoffrey Hinton, often called the godfather of AI technology, have warned that AI technology could lead to anyone soon having the power to create weapons of mass destruction. Erik Brynjolfsson, one of the world's leading AI experts, recently said: "If we continue down this path, the risk of an extreme concentration of power, both economic and political, is very real."
"I am appalled by the lack of commitment in universities, among decision-makers, think tanks and parts of the private sector"
In August, three engineers at Anthropic, one of the world's leading AI companies, made headlines when they quit the company on the grounds that the technology could become an existential threat to humanity in the future.
Bill Gates is now trying, as humbly as he can, to remind that for almost 50 years he has been at the center of American technological innovation and therefore has some weight when he now talks about the risks.
So what specifically is he worried about AI doing?
Gates mentions four areas where AI is about to reach "critical thresholds," meaning the technology is becoming so powerful that it will be difficult to stop. This applies to cyber attacks, bioterrorism, the effects on the labor market and what he calls "psychosocial" effects - how AI affects people's social lives and mental state.
- In the end, it's about us losing control over the systems themselves. I am appalled by the lack of commitment to this in universities, among decision makers, think tanks and parts of the private sector. Therefore, I felt that I had to step forward and speak clearly that AI is not similar to previous technologies.
Bill Gates turned 70 last year, but it was a bittersweet celebration that coincided with a turbulent time in his private life. He went through a divorce after more than 30 years of marriage with Melinda Gates, with whom he also ran his large foundation. He was also called for questioning in the US Congress to clarify business relations with the sex offender Jeffrey Epstein. Gates claims their meetings were about Epstein, who was very wealthy, promising to donate money to Gates' foundation. But the promises were never fulfilled; instead, Epstein tried to blackmail Gates about an affair.
It was a year of challenges – events that surely affect how people view my credibility as a messenger
The interrogation did not lead to any further charges against Gates, he is not suspected of any crime, but he says that the turmoil has affected his own priorities going forward.
- It was a year of challenges - events that certainly affect how people see my credibility as a messenger.
He continues:
- But I think we will need as many voices as possible warning about this. None of them will be a perfect person. I know many who lead the AI industry so I might be expected to be more cautious about criticizing them. But this is the first time in my life I've done it.
His cautious voice suddenly takes on a certain pathos, as if he were addressing an audience.
Gates says he is "50 times more worried" that AI could lead to bioterrorism and social chaos than he is about another pandemic on the scale of Covid, which killed more than seven million people.
- When people say: "Oh, we handled personal computers and the internet, won't it be the same now?" then I answer no, it won't be the same. It's different this time. Both the benefits and the risks are of a completely unique nature, unlike anything else. So I hope we get people on board, because I see the need to speed up the positive and minimize the destructive as extremely urgent.
"We have a chance to act now, so that unemployment does not skyrocket and public confidence disappears"
In the short term, he is most concerned about the effects on the labor market. Even conservative forecasts expect at least 7-10 percent of office jobs to disappear in the near future. Gates says AI could become “the biggest factor of inequality or of equality” in our lifetime, depending on how the coming years play out politically.
Large sections of the American people share the concern. 79 percent of Americans believe that the AI industry could endanger their jobs and 72 percent believe that AI companies will push down their wages, according to The New York Times.
- We have a chance to act now, so that unemployment does not skyrocket and public trust disappears, he says.
So what can we do concretely? Gates first of all wants to dispel the myth that companies will regulate themselves. Both Donald Trump and the Speaker of the US House of Representatives, Mike Johnson, have consistently argued that we must only rely on companies' own ability to pull the brakes when necessary.
Gates makes it clear that is unrealistic.
- I don't think you can just say to an industry: "Okay, solve the regulation yourself." The AI companies themselves have said that they believe they need to be regulated. Their sincerity will be put to the test once we present some of the concrete proposals. Some of my safeguards for areas like biosecurity would barely slow the AI market at all.
I never thought it was realistic to try to 'pause' development, but wanted to see comprehensive regulations and brakes on the risks
Bill Gates no longer has a formal role at Microsoft, where he left as CEO more than twenty years ago. He left the board in 2020, but he has regular conversations with both Microsoft and the other tech giants' executives, including Open AI. With the Gates Foundation, he also has business connections with Anthropic and Google.
In the technology journalist Karen Hao's book "Empire of AI", she describes how Bill Gates was initially skeptical of Open AI's development of generative AI, which he first saw back in 2019. When Microsoft then wanted to invest in the company, Gates is said to have warned CEO Satya Nadella that they were "wasting a billion".
But three years later, Open AI's leadership was invited to Gates' home for a dinner where they showed a demonstration of Chat GPT, which Gates described as the most awe-inspiring moment he's experienced in 40 years of technological development. He encouraged a giant investment by Microsoft, which soon after invested 12 billion dollars in Open AI.
The dinner meant that Gates himself became a key person in the acceleration of the very technology he is now warning about. When he saw the potential, his instinct was to encourage Microsoft to pump money into the company—not to pull the handbrake.
This is not necessarily a contradiction, he now believes.
- I never thought it was realistic to try to "pause" development, but wanted to see comprehensive regulations and brakes for the risks. Originally, they wanted to limit the damage based on the assumption that only one company was at the forefront: Open AI. But today we have a situation with several companies in several countries and then we find ourselves in a situation where it is not enough for the industry itself to hold back. It will take governments.
The major threats, such as bioterrorism, are global and instead of a competition between countries, cooperation between the US and China is absolutely necessary
What he means is that international agreements are needed, in the same way as we had for nuclear weapons, air traffic or emissions. The risk is that such a process drags on for time, and in the meantime the business leaders can safely sit at conferences and talk about how nice it is with regulations, without having to make it concrete, because it will not happen under sitting governments.
Something else that is vital but currently unrealistic is an improved cooperation between China and the US on the security issues surrounding AI.
Bill Gates himself can become a mediator between the two countries, as he is one of the few people in the world who has good relations with both Washington and Beijing.
- The major threats, such as bioterrorism, are global and instead of a competition between the countries, cooperation between the US and China is absolutely necessary.
He will meet with Xi Jinping later this year to discuss this, and has already met with the Trump administration several times.
Does Trump understand what's at stake?
- I don't think anyone outside the industry fully understands. Some see only the positive, others only the negative. I would say that the current US administration has mostly taken the attitude of: "Go ahead, clear all obstacles", and I would like to see a better balance. Governments will need to increase the competence of the state apparatus and step up the dialogue with the industry, because the industry really understands the cyber risks and many other things on a very deep level.
Gates is helpfully specific about the taxes and regulations he wants to see here and now. He is advocating a new tax on robots that goes to retraining those who lose their jobs. He also wants to see significantly higher corporate taxes in the US, something you don't hear every day from American tech moguls. Billionaires like Elon Musk, Larry Ellison and Jeff Bezos often rail against similar proposals. Google founder Sergey Brin has poured about $100 million in private money into stopping a marginal increase in California's wealth tax, which he claims would "turn California into the Soviet Union."
Bill Gates, on the other hand, has promised to give away basically his entire fortune, through his foundation, the Gates Foundation. During the interview, he emphasizes that AI technology can bring about enormous progress in medical research, cancer diagnoses, climate research, energy supply, and not least aid work, which he now spends most of his working time on.
He gets most involved when we talk about this, especially the trips to African countries where the foundation is most active. A tragic consequence of the moral vacuum in our time's political and economic leadership is that rich countries have sharply cut back on humanitarian aid. This applies not only to the USA, but also to the large countries in Europe. The consequences can be catastrophic. According to the American The Lancet, 14 million may die due to the US cuts alone.
I was in Rwanda and had a long dinner with President Kagame, where we discussed how AI can improve healthcare on the ground
- It is really important to highlight the negative consequences of the fact that - not only in the USA, but quite broadly in Europe - aid has been de-prioritised, even though it constitutes a vanishingly small part of the budget. People simply do not know what the aid work has achieved. That child mortality has been halved, that we are on the way to eradicating malaria.
In aid work, he sees how AI can also be used to improve and streamline humanitarian work in the most vulnerable countries.
- I was in Rwanda a couple of weeks ago and had a long dinner with President Kagame, where we discussed how AI can optimize the way we work with aid, track the money and improve healthcare on the ground.
So AI can be both the cause and the solution to the problems?
- We have no choice but to improve our work, because the pressure on state budgets - whether it is about an aging population, increased defense spending or investments in safety nets to mitigate the effects of AI - will compete directly with aid.
For Gates' foundation, that means more urgency to get rid of the money they're sitting on. Recently, he announced that the foundation will close in 2045. He does not want to create a permanent institution, but make sure that the fortune is spent now when it is really needed.
- The foundation now only has 19 years left. We need to create new partnerships. We have strong collaborations in Europe and especially in Sweden. But if we can't gather a large crowd of voices, both on the importance of aid and the importance of controlling AI, I'm really afraid of the damage.
- So just see it as a warning, from a messenger who is human.
