Geo News
Community curated by people like you
LatestAICryptoHealthWorld AffairsUS Politics
Demis Hassabis exits as DeepMind CEO after pitching AI safety oversight body
00

Demis Hassabis exits as DeepMind CEO after pitching AI safety oversight body

Aug 12, 2026

Demis Hassabis has stepped down as CEO of Google DeepMind to become chair and Alphabet's chief scientist, following a major restructuring on August 5, 2026. The transition occurs as Google's Gemini model faces development delays and misses critical deadlines, trailing rivals like Anthropic and OpenAI. Google co-founder Sergey Brin has stepped in informally to pressure researchers and steer resources toward self-improving AI, while Hassabis had spent his final weeks as CEO pitching an independent global AGI safety watchdog modeled after the IAEA.

Hassabis DeepMind CEO exit

  • ▪The Pathfounders newsletter reported that Demis Hassabis wanted to exit Google entirely alongside Jeff Dean but was persuaded to accept the chair role to prevent an Alphabet share price collapse.
  • ▪Koray Kavukcuoglu took over day-to-day control of Google DeepMind as chief executive officer following the August 5, 2026 leadership transition.
  • ▪Demis Hassabis stepped aside as chief executive officer of Google DeepMind to become chair and Alphabet chief scientist on August 5, 2026.

Gemini performance lag

  • ▪Google DeepMind engineers traced Gemini's coding gap to decisions made in mid-2025, when Google focused on protecting search traffic while competitors advanced AI agents.
  • ▪Google delayed the release of its flagship Gemini model by two months after internal testing showed it trailed competitors in coding capabilities.
  • ▪Google's Gemini development faced delays due to constrained computing capacity, disagreements among project leads, and bureaucratic release cycles.
  • ▪Google's Gemini 3.5 Pro model missed three release deadlines after being unveiled at Google I/O in May 2026 with a promise to ship the following month.

DeepMind organizational restructuring

  • ▪Gemini technical co-leads Jeff Dean and Oriol Vinyals left Google DeepMind during the August 2026 shakeup to launch Discovery Loop, a Google-backed startup focused on self-improving AI.
  • ▪Google announced a sweeping reorganization of its artificial intelligence division on August 5, 2026, which shifted several nontechnical teams out of Google DeepMind and into Google's corporate reporting structure.
  • ▪Alphabet shares fell approximately 5% following the announcement of the Google DeepMind reorganization and leadership changes.

Brin informal AI influence

  • ▪Sergey Brin has used his informal influence to direct Google DeepMind resources toward recursive self-improvement, where artificial intelligence models can improve without human intervention.
  • ▪Google co-founder Sergey Brin has been informally pressing Google DeepMind researchers to accelerate Gemini development to close the gap with competitors like Anthropic and OpenAI.

AI safety oversight proposal

  • ▪In the weeks before stepping down as Google DeepMind chief executive officer, Demis Hassabis pitched government officials and industry leaders on forming an independent artificial general intelligence safety watchdog.
  • ▪Demis Hassabis likened his proposed independent artificial general intelligence safety entity to the International Atomic Energy Agency.

2 sources

Timesofindia
Google co-founder Sergey Brin has a message for DeepMind staff after Demis Hassabis exits as CEO: Catch up to the...
View source article
Wsj
Exclusive | DeepMind’s Hassabis Pitched AI-Oversight Body Before Shake-Up
View source article

Featured stories

View more in AI governance

Trump signs voluntary AI safety accord with tech executives

Sep 29, 2026 · 14 sources

Anthropic warns of existential risks to humanity in IPO prospectus

Sep 28, 2026 · 6 sources

RSA launches Agent ID security platform to track thousands of shadow AI agents in enterprises

Sep 28, 2026 · 3 sources

Florida seeks court order to halt OpenAI model development

Sep 28, 2026 · 4 sources

Story comments

Loading comments…

People Involved

Demis HassabisSergey Brin

Related Projects

DeepMindGoogleGeminiAlphabet

Topics

AI governanceAI standards, audits & complianceAI safety & social impact

Featured stories

View more in AI governance

Trump signs voluntary AI safety accord with tech executives

Sep 29, 2026 · 14 sources

Anthropic warns of existential risks to humanity in IPO prospectus

Sep 28, 2026 · 6 sources

RSA launches Agent ID security platform to track thousands of shadow AI agents in enterprises

Sep 28, 2026 · 3 sources

Florida seeks court order to halt OpenAI model development

Sep 28, 2026 · 4 sources