Google DeepMind CEO Wants AI Cops to Babysit Rogue Thinking Sand
KEY POINTS
- •Demis Hassabis proposed a U.S. AI watchdog funded by the tech industry to screen advanced models before deployment.
- •He briefed the Trump administration and European officials while advocating for swift implementation by the end of 2026.
- •The crackdown on Anthropic’s Mythos models highlighted the need for formal regulations beyond ad hoc government orders.
Demis Hassabis, Google DeepMind co-founder, Nobel laureate, and part-time AI whisperer based in London, is pushing for a U.S. AI watchdog before his company’s Gemini brainchild melts down civilization. Published on May 26, 2026, his manifesto—'A Framework for Frontier AI and the Dawning of a New Age'—calls for a FINRA-style regulator funded by Big Tech to babysit AI models 30 days pre-launch like nervous parents at a science fair. After briefing everyone from the Trump administration (yes, they went from AI laissez-faire to 'Mythos scare' mode) to European officials, Hassabis wants this watchdog up and running by year-end. His pitch: screen AI threats from cyber nightmares to nuclear-level horrors lurking in open-source sand brain models only he’s willing to admit are about 'to think.' Also, Anthropic’s CEOs joined the choir for binding rules after getting their Mythos AI frozen overnight like bad TikTok trends.
Share the Story
(1 of 3)When your AI can cause nuclear apocalypse but industry watchdogs are modeled after Wall Street referees, you know the robot apocalypse might have a compliance form to fill first.
AI now comes with a 30-day testing period where tech giants nervously pray it doesn’t hack your toaster into a missile launcher.
Had enough of government regulations? Don’t worry, now AI's being babysat by FINRA fans who gave us Wall Street chaos, so what could go wrong?
Source: Axios | Published: 7/14/2026 | Author: Mike Allen
More Articles in Technology
Digital Nomad Trapped in Lisbon FOMO, Ends Up BFF Ghosted
Senate Passes Russia Sanctions Act, Names It After Dead Senator To Speed Up Votes
Sam Altman Throws Shade So Hard, Even AI Feels Betrayed
Wix CEO Admits AI Is Less Terminator, More Clumsy Intern
Microsoft Cleans Up Windows Search by Finally Giving Up on Selling Ads
OpenAI CEO Welcomes Criticism, Employees Insist They Haven't Been Fired Yet
Palantir CEO Warns AI Will Make Billionaires Richer and Everyone Else Miserable
