Hassabis Pitched IAEA-Style AI Safety Body Before DeepMind Leadership Shake-Up
Summary
Updated Aug 13 — Hassabis pitched the IAEA-style AI safety body specifically to Trump administration officials, including Treasury Secretary Scott Bessent.
- • Demis Hassabis pitched an independent AI safety entity modeled on the IAEA to governments and rival lab leaders before stepping down as Google DeepMind CEO
- • Hassabis proposed the body to codify safety guardrails and best practices for AGI development, holding talks with government officials and other AI labs
- • The 'shake-up': Hassabis became Chairman of Google DeepMind and Chief Scientist of Alphabet around August 5–6, 2026
- • Koray Kavukcuoglu took on responsibility for AI-model development at Google DeepMind following the restructuring
Updates
Pitch specifically targeted Trump officials including Bessent
According to WSJ sources, Hassabis's pre-departure outreach specifically included top Trump administration officials; Treasury Secretary Scott Bessent is named as one of the senior officials who received the AI safety body proposal — adding political specificity to earlier reporting that described outreach to 'government officials' broadly.
Details
IAEA-style AI safety body proposed
Hassabis proposed an independent entity to codify safety guardrails and best practices for AGI development, explicitly likening it to the International Atomic Energy Agency in discussions with peers.
Outreach to governments and rival labs
In the weeks before stepping down as CEO, Hassabis discussed the proposed safety body with government officials and leaders from other AI laboratories.
Hassabis steps down as DeepMind CEO
Around August 5–6, 2026, Hassabis relinquished his role as CEO of Google DeepMind, becoming Chairman of Google DeepMind and Chief Scientist of Alphabet.
Kavukcuoglu leads AI model development
Koray Kavukcuoglu took on responsibility for AI-model development at Google DeepMind as part of the restructuring.
Safety work as Hassabis's new focus
Hassabis has been spending increasing time advancing the safety body concept, signaling a deliberate strategic shift toward policy and governance over product management.
Sources: The Wall Street Journal (exclusive); WSJ / NYT on leadership changes, August 2026; details corroborated via Grok live web research.
What This Means
Hassabis pitching an IAEA-style AI oversight organization represents a significant moment in AI governance: a leading AI scientist attempting to establish international coordination structures for AGI before those systems arrive. His transition from hands-on CEO to Alphabet's top scientist may reflect a deliberate choice to prioritize cross-industry safety work over product management, giving him the standing to engage governments and peer labs as a statesman rather than a competitor. The identification of specific Trump administration officials — including Treasury Secretary Scott Bessent — as recipients of the pitch reveals that Hassabis was targeting the highest levels of the current US government, not just the broader policy community. The proposal — if it gains traction — could become a template for how the industry self-organizes around existential risk, or the opening move in a longer fight over who controls AI oversight.