Bill Gates Calls for a Global Action Plan on AI Risks
The Microsoft co-founder describes the shift to AI as one of history's most turbulent phases and wants oversight bodies instead of industry self-policing.
Illustration: a deserted port terminal at blue hour where automated cranes and robotic arms stand motionless.
Bill Gates wants a globally coordinated action plan for AI because, in his view, the world is not prepared for the disruption already in sight.
At a glance
- Gates made his case on August 26, 2026, in a GatesNotes essay and an interview with the New York Times.
- He highlights three risks: permanent job losses, biological misuse and psychological harm to users.
- The most exposed roles, he says, are entry- and mid-level jobs in sales, customer service, software development and legal support.
- He wants international oversight bodies modeled on the regimes built around nuclear weapons.
- His proposals include a levy on AI usage and robots, plus care occupations reserved for human workers.
Bill Gates is calling for a globally coordinated action plan on artificial intelligence, arguing that the world is not ready for the upheaval the technology is about to cause. He set out the case on August 26, 2026, in an essay on his GatesNotes blog and in an interview with the New York Times.
A warning aimed at his own industry
Gates describes the move into the AI era as one of the most turbulent phases in human history. His sharpest criticism lands on technology companies, which he accuses of playing down risks they already understand. Leaving the industry to police itself, he argues, will not be enough.
The three risks he singles out
The first is permanent job loss rather than a temporary adjustment. Entry-level and mid-level roles are the most exposed: sales, customer service, software development and legal support. A Stanford study cited in the coverage finds employment among young workers in the most exposed occupations falling since generative AI became widely available.
The second is biological misuse. Gates argues that advanced models lower the barrier to designing new pathogens far enough that a small group could attempt it. The same capability that helps defenders close security holes also helps attackers find them.
The third is psychological. Research from Stanford and Carnegie Mellon University found that the more heavily people leaned on emotionally oriented chatbots, the worse they reported feeling. Gates is also worried about children, and about the erosion of critical thinking at a moment when convincing deepfakes are cheap to produce.
What he proposes
For frontier systems, Gates wants international oversight bodies modeled on the arrangements built around nuclear weapons. On the economic side he floats a levy on AI usage and on robots, described in the coverage as a token tax. He pairs it with a “human reserved” idea: keeping certain care and support occupations explicitly staffed by people.
What the available sources do not settle
The sources available for this article give no detail on how such a levy would be assessed, at what rate, or by which authority. The Golem.de report was reachable only through its cookie consent screen at the time of research, so only the headline and summary could be used. A Zeit Online piece on the same subject could not be retrieved at all and is therefore not listed as a source.
FAQ
What is Bill Gates asking governments to do about AI?
He wants a globally coordinated action plan, including international oversight bodies for advanced AI systems, and explicitly rejects letting the industry regulate itself.
Which jobs does Bill Gates say AI will hit first?
Entry-level and mid-level positions in sales, customer service, software development and legal support, where he expects the losses to be permanent rather than temporary.
What is Bill Gates' proposed tax on AI and robots?
A levy on AI usage and on robots, referred to in the coverage as a token tax. How it would be calculated or collected is not specified in the available sources.
