Hassabis pitched AI safety body to Trump officials before DeepMind exit

The Wall Street Journal reported on 13 August 2026 that Demis Hassabis pitched senior Trump administration officials on a new independent AI safety body modelled on the International Atomic Energy Agency, in the weeks before he stepped down as Google DeepMind chief executive.

The Wall Street Journal said Hassabis raised the proposal with Treasury Secretary Scott Bessent and science adviser Michael Kratsios, as well as with other frontier AI companies.

The talks took place before Hassabis gave up the DeepMind CEO role on 5 August, according to the report. He now serves as DeepMind chair and Alphabet chief scientist.

The proposed body would set safety guardrails and best practices for developing artificial general intelligence, the report said. Hassabis has been spending more of his time on the idea, according to the Journal.

Hassabis had already called publicly for a US-led standards body. On 14 July he told CNBC that a public-private group, overseen by the federal government, should test new frontier models for national-security risks, according to CNBC. In June, Hassabis and Anthropic's Dario Amodei urged a US-led coalition on AI rules at a G7 meeting.

His broader plan envisions national AI safety institutes, similar to the UK's, feeding into a global body that would set safety benchmarks, certification requirements and independent audits of frontier systems, according to the Wall Street Journal report summarised by OODA Loop.

Hassabis said earlier this year on the 20VC podcast: "Nobody should be building systems that are capable of deception because then they could be getting around other safeguards."

United Nations Secretary-General António Guterres endorsed the idea of an "IAEA for AI" in 2023. After agent-related incidents at OpenAI and Anthropic this summer, labs have pushed for common safety standards, and a group of companies later began discussing an industry-led standards body.