DeepMind’s Demis Hassabis Pushes US‑Led AI Watchdog as Race for ‘Frontier’ Models Accelerates

DeepMind’s Demis Hassabis Pushes US‑Led AI Watchdog as Race for ‘Frontier’ Models Accelerates
Google DeepMind chief Demis Hassabis is escalating calls for a US‑anchored global AI watchdog just as competition to deploy ever more powerful “frontier” models intensifies and formal regulation still lags.

Early warnings and a “precious window”

Hassabis, a Nobel laureate in chemistry for his work on AI‑driven protein prediction, has warned that society has only a limited window to prepare for technology advancing at “historic speed.” He argues that artificial general intelligence could arrive “only a few short years away,” framing the current moment as “the foothills of the singularity.”

July 14: Blueprint for a global AI watchdog

On July 14, Hassabis laid out his most detailed plan yet, arguing that “it’s time for a global AI watchdog” with power to “hit the brakes” if frontier models become too dangerous. He proposes a new standards body, ideally led by the US “given its economic and technical standing,” to test cutting‑edge systems before release and coordinate an industry‑wide slowdown when risks are too high.

A companion proposal envisions an independent “standards body” modeled on Wall Street’s FINRA that would test models up to 30 days before deployment and develop best‑practice guidelines for their release. Initially voluntary, this assessment could quickly become mandatory for any model entering the US market once proven effective.

Balancing US politics, industry control, and global reach

Hassabis has spent months quietly briefing the Trump administration, other AI labs, and European officials, saying feedback from Washington has been “very positive.” Positioning the body as a self‑regulatory organization funded by AI companies but backed by the US government is also meant to answer skeptics in the White House who insist “there will not be an FDA for AI.”

Staffed by independent technical experts and open‑source representatives and able to outsource specialist risk assessments to AI safety groups, the watchdog is pitched as “technically focused” while still “supporting innovation and incentivising” progress, rather than halting it. For now, it remains a proposal—but one Hassabis hopes to see operating before year’s end.

Continue reading https://foxvector.com/stories/019f6613-a624-0414-70d2-0b98b76420cf