Bill Gates warns of the danger of artificial intelligence in enabling devastating cyber and biological attacks

He said that the ability to launch a cyber attack that destroys all bank accounts and disrupts the electricity grid “also exists today, and we know that is the case.” He attributed this to the fact that "anyone with malicious intentions can use artificial intelligence and force him to carry out these actions."

Gates added that awareness of AI risks outside the industry is “extremely low.” He explained that this is true for academic circles, think tanks, policy makers, and politicians alike.

He noted that part of the reason he was speaking out "as loudly as I could" was that "this industry can't be counted on to self-regulate on this. I mean, this is crazy."

Gates asked: “Should we wait until a cyber attack causes serious damage, and a biological attack causes serious damage, to then impose oversight guarantees in these models to limit the possibility of this happening again and again? Or can we be wise enough to put these guarantees in place, and be obligated to put them in place, before millions are killed?”

Gates' statements came on "The Ezra Klein Show", where he said that artificial intelligence crossed the cyber and biological thresholds early this year. He cited models that discover security vulnerabilities in codes that have been reviewed by humans for more than 20 years, adding that the ability to design a new pathogen was limited to countries, and is now within the reach of small groups.

Gates called for a supervisory layer of safeguards and oversight in all AI models. He suggested keeping the models open source, free and customizable, as long as they ran on platforms that allow governments to verify that safeguards remain. On the other hand, he believed that Anthropic had set the filters of its “Mythos” and “Vible” models at too high a level.

Gates had previously warned, on NBC's "Meet the Press" program, that artificial intelligence could cause a billion deaths.