Microsoft co-founder Bill Gates is issuing an unusually stark warning about the potential dangers of artificial intelligence, saying advanced AI tools in the hands of malicious actors could set off a chain of events resulting in catastrophic loss of life.
In an interview broadcast Sunday on NBC's Meet the Press, Gates told host Kristen Welker that AI is already powerful enough to potentially facilitate events resulting in as many as one billion deaths.
"AI is definitely powerful enough to drive events that could cause a billion deaths," Gates said.
He then explained what concerns him most: the combination of increasingly powerful AI systems and people intent on using them for destructive purposes.
"There has never been a weapon as powerful as the combination of malicious people using the latest AI tools," Gates said.
Gates' reference to one billion deaths was not a prediction that AI will actually kill that many people. Rather, he was describing the potential scale of a catastrophic scenario in which powerful AI technology is deliberately misused.
His broader message was that responsibility for AI safety cannot be left solely to the technology companies developing these systems.
Gates called for direct government involvement in regulating and overseeing AI, arguing that lawmakers and US law enforcement agencies must have a role in determining safeguards for the rapidly evolving technology.
"Nobody thinks self-regulation is enough," he said.
Gates acknowledged that mandatory oversight would impose additional costs on the AI industry, but argued that regulation should be incorporated into the development process rather than left to individual companies to decide voluntarily. He said such measures need not dramatically slow the industry's progress.
His latest warning comes amid growing concern over the risks associated with the rapid development of AI.
Earlier this month, Gates told Reuters that governments around the world are falling behind in preparing for changes expected to be brought about by AI. He pointed to potential risks involving employment, cyberattacks and addictive AI tools.
At the same time, Gates continues to emphasize AI's enormous potential to improve people's lives, particularly in healthcare, education and agriculture.
The threat posed by malicious use of advanced AI is already more than a theoretical concern. Earlier this month, Anthropic said it had disrupted attempts to use its Claude AI models for dangerous purposes, including biological weapons research and cyber operations.
According to the company, one case involved a user seeking assistance in planning experiments involving avian influenza, while other cases involved cyber espionage.
The risks posed by AI have also become an issue in relations between the world's major powers. The US and China have agreed to establish a formal dialogue on AI risks and emergency communications. Discussions have included concerns about AI agents potentially operating beyond human control, as well as threats from non-state actors, including cyberattacks.
Gates' warning comes as debate within the technology industry over the limits of AI development grows increasingly intense. While some are calling for stronger safeguards, there is still significant disagreement over the scale of the risks and how they should be addressed as the global race to develop increasingly advanced AI systems accelerates.
Gates is not calling for AI development to be halted. His message is more nuanced: The same power that could enable AI to transform healthcare, education and science could also make it extraordinarily dangerous when placed in the hands of people seeking to cause harm.
The greatest vulnerability, in his view, may therefore not lie in the machines themselves, but in the combination of increasingly powerful technology and humans determined to misuse it.
And to underscore just how catastrophic that risk could potentially become, Gates chose a number that is almost impossible to ignore: "one billion people."