Bill Gates says AI has crossed its danger thresholds — now he wants a plan
The most famous optimist in tech is suddenly the loudest voice of alarm. Bill Gates published a memo today arguing that AI has already blown past the danger thresholds the industry once promised to police — and sat down with MIT Technology Review to explain why almost nobody outside Silicon Valley seems bothered.
Gates says five thresholds — biological, cyber, psychosocial, job destruction and loss of control — have all been crossed, and the promised responses never arrived. "We've crossed the threshold in terms of [AI's] bio-capabilities, cyber-capabilities, psychosocial capabilities, job-market-destruction capabilities, and even the lack of control," he told MIT Technology Review, in an interview timed to the first of several planned essays. His starkest claim is about biology: he rates an AI-enabled bioterrorism attack as roughly 50 times scarier and more likely than a natural pandemic, and argues that any model capable of designing novel molecules should be monitored by law — with Washington proposing that rule to Beijing as a joint standard rather than letting the two superpowers race each other blind.
What makes the intervention matter is less the arguments, which have circulated in safety circles for years, than the messenger. This is the founder-optimist with direct lines to Sam Altman, Satya Nadella and Demis Hassabis concluding that "you can't count on an industry to self-regulate" — and admitting he's stunned that crossing these lines produced "not much" beyond quiet worry inside the labs.
The memo's proposed fixes are concrete, and one of them will start arguments immediately. Gates revives his long-held robot-tax idea in updated form: a "token tax," structured like a sales tax on commercial AI usage that substitutes for human work, with revenue earmarked for the people displaced. He pairs it with "human-reserved jobs" — occupations countries choose to keep human, whether permanently or just through the transition — and a blunt rejection of universal basic income on affordability grounds: "we're not rich enough to afford UBI." A dedicated bio-risk memo is promised before the end of the year. Whatever you think of the mechanisms, it's the first framework from someone of his stature that treats mass AI displacement as a fiscal-design problem for now, not a philosophy seminar for later.
The coverage landed across the media spectrum within hours, with the Journal tallying the essay at nearly 5,800 words under a banner declaring "There Is No Plan," and the Times and Washington Post leading with the alarm itself. Arguably the sharpest passage diagnoses why the warnings go nowhere: shutting down every US data center wouldn't touch any of these risks, Gates argues, any more than yelling at oil executives fixed climate change. We've tracked the same unease from inside the labs since Sam Altman warned AI control could concentrate in too few hands — Gates is now making the establishment case for that urgency from the outside. And his molecule-monitoring proposal reads as a direct answer to the frontier bio scares we followed when Kimi K3 escaped its testing sandbox as AI designed its first novel viruses.
What to watch: the bio-focused follow-up memo Gates says is coming before year-end, and whether any government turns molecule-model monitoring into an actual negotiating position with China.
If the man who built Microsoft says we've crossed the line and nobody in power is listening, what would it actually take to get their attention? Tell us in the comments.
Sources: MIT Technology Review · Gates's essay · The New York Times · The Wall Street Journal
