Hassabis backs Amodei’s slowdown call and points to a DeepMind standards body
The DeepMind chief said the direction is correct and the details still need work. Altman and Musk already agreed this morning. Four lab heads is not a pause.
Feature1 minIn-depth
Maya Chen1 min read
- DeepMind
- Anthropic
- Hassabis
- Safety

Photo: Anadolu AgencySource page
Demis Hassabis, who runs Google DeepMind and is Alphabet’s chief scientist, publicly backed Dario Amodei’s call to pace frontier development on Sunday. Anadolu Agency, Moneycontrol and New Kerala carried the remarks. On X, Hassabis wrote that Amodei’s essay “points towards the right path forward.” He added that the details still need working through, “but the direction is correct for meeting this critical moment.”
He tied the comment to a recent DeepMind proposal for an industry-wide standards body for frontier AI. That is a different first step from Anthropic’s: Amodei committed to giving independent evaluators permanent, employee-level access. Hassabis is arguing for shared rules, not yet for badges inside Google’s labs.
Dario's essay points towards the right path forward. The details need working through, but the direction is correct for meeting this critical moment. This is also why we recently put out our proposal for an industry-wide standards body for frontier AI.
— Demis Hassabis, post on X, as quoted by Moneycontrol and New Kerala, Sep 13, 2026
Sam Altman and Elon Musk had already agreed with Amodei earlier on Sunday. Altman said OpenAI will match the evaluator commitment; Musk wrote “Dario is right.” Four named lab heads now sit on the same side of a sentence. None of them announced a delayed training run, a slipped ship date, or a joint evaluation protocol. New Kerala also recorded former UK prime minister Rishi Sunak calling for stronger independent verification, and US senator Bernie Sanders repeating a demand to pause advanced development and ban artificial superintelligence.
Amodei’s essay, posted Saturday, asked labs to slow the rate of capability gains so alignment and evaluation can keep up. Anadolu quoted him: “We must slow the pace at which we improve the capabilities of AI models. Progress will still seem fast, and we must make wise use of the time we gain.” Hassabis agreeing on direction, and pointing at a standards body, leaves the working-through to the week ahead.


