Demis Hassabis is a co-founder of DeepMind and a 2024 Nobel laureate in Chemistry, shared with John Jumper and David Baker, for AlphaFold. He led Google DeepMind as chief executive from the 2023 merger until August 2026, when he became chair of the unit and chief scientist of Alphabet. A former chess prodigy with a neuroscience PhD from UCL, he is among the more scientifically credentialed figures in frontier AI leadership and has emphasized AI's application to scientific problems.
Role: Chair, Google DeepMind, and chief scientist, Alphabet (from August 2026); CEO, Google DeepMind (2014–2026; merged entity since 2023) Previous: Co-founder, DeepMind Technologies (2010); chess prodigy, neuroscience PhD (UCL) Honors: Nobel Prize in Chemistry (2024, with John Jumper and David Baker) for AlphaFold; Knighthood (2024)
Background
Hassabis co-founded DeepMind in London in 2010 with Shane Legg and Mustafa Suleyman (now at Microsoft AI). Google acquired DeepMind in 2014. In 2023 Google merged DeepMind with Google Brain into Google DeepMind under Hassabis's leadership, consolidating Google's AI efforts under a research-led rather than product-led structure. His research background spans game AI (AlphaGo, AlphaZero), protein structure prediction (AlphaFold 2 and 3), and scientific computing (GNoME, materials discovery). He is a corresponding author, with David Silver, on the January 2016 Nature paper describing AlphaGo (Mastering the Game of Go with Deep Neural Networks and Tree Search (Silver et al., Nature, January 2016)).
Roles and work
Hassabis led AlphaFold 2 (2020) and AlphaFold 3 (2024), the first AI systems widely adopted as primary tools in structural biology; the work was cited by the 2024 Nobel committee. He has led Gemini development across generations, most recently Gemini 3.
On August 5, 2026 Sundar Pichai told Google DeepMind teams that Hassabis would move from chief executive of the unit to chair of Google DeepMind and chief scientist of Alphabet, with Koray Kavukcuoglu — previously the unit's chief technology officer and Google's chief AI architect — taking over daily operations as senior vice president. Hassabis wrote that "it's critical that we collectively get the next steps right to ensure this all goes well for humanity," said he would spend more time at Isomorphic Labs, and stated that he has "always believed the No.1 application of AI should be to improve human health." He named the unreleased Gemini 4 in describing the company's progress (Source: blog.google; fortune.com).
In September 2026 Hassabis became one of three directors of the DeepMind Institute, a publishing forum on AGI launched by Google and Google DeepMind on September 16, with Shane Legg as managing editor and James Manyika (Source: axios.com). The three directors' launch essay states that AI is approaching AGI, "a system that exhibits all the cognitive capabilities of the human brain," and that "we expect those gaps to be closed soon"; the Institute's biography of Hassabis lists him as Nobel laureate, co-founder and chair of Google DeepMind, and chief scientist of Alphabet (Introducing the DeepMind Institute (Legg, Manyika and Hassabis, September 2026)). His July 2026 framework essay (A Framework for Frontier AI and the Dawning of a New Age (Hassabis, July 2026)) is one of the Institute's five inaugural publications (Source: thenextweb.com).
In 2021 he founded Isomorphic Labs to apply AlphaFold-derived methods to drug discovery, and remains CEO of Isomorphic. He told the Financial Times in 2024 that Isomorphic would have AI-designed drugs in clinical trials by the end of 2025; that target slipped. President Max Jaderberg said at WIRED Health on April 16, 2026 that the company is "gearing up to go into the clinic" but did not give a hard date. (Source: wired.com)
Positions and statements
Safety and risk
Hassabis signed the 2023 CAIS Statement on AI Risk ("Mitigating the risk of extinction from AI should be a global priority"). He has framed AGI as likely within the decade while advocating cautious deployment. Google DeepMind publishes a Frontier Safety Framework analogous to OpenAI's Preparedness Framework and Anthropic's RSP. He signed the Seoul Frontier AI Safety Commitments.
Hassabis endorsed Amodei's pacing essay on the evening of September 12, 2026: "Dario's essay points towards the right path forward. The details need working through, but the direction is correct for meeting this critical moment." He did not specify a pace or an oversight arrangement (Source: axios.com).
Regulation
Hassabis has generally supported government-led safety evaluations and the UK AISI / US AISI network. He participated in the UK's Bletchley Park summit producing the Bletchley Declaration (2023). He has been less publicly combative on state regulation than Meta or xAI leadership.
On July 14, 2026, Hassabis published "A Framework for Frontier AI and the Dawning of a New Age," proposing a U.S. AI standards body modeled on FINRA, the financial industry's self-regulatory organization. Under the proposal, frontier labs would initially share their models voluntarily up to 30 days before release for safety testing focused on dangerous cyber, biological, and deception capabilities; passage through that testing would later become mandatory for U.S.-market deployment of all frontier-class models, open or closed, regardless of country of origin (A Framework for Frontier AI and the Dawning of a New Age (Hassabis, July 2026); Source: axios.com). Hassabis said he had briefed the Trump administration, fellow lab leaders, and European officials on the proposal, and that he wanted the body — industry-funded, with a majority-independent board — operational "before year-end" (Source: theverge.com). He called the administration's June 2026 export-control freeze of Anthropic's Mythos and Fable models "a bit of a wake-up call" (Source: axios.com). See AI Pre-Release Vetting. Gary Marcus welcomed the proposal the same day while cautioning that transparency and independence remain essential given U.S. government interest in taking stakes in AI companies (Source: garymarcus.substack.com). By July 17, 2026 the proposal — described in coverage as a Frontier AI Standards Body that would audit frontier models for cybersecurity and biological-threat capabilities before release — had drawn public endorsements from Sam Altman and Elon Musk, and praise from Anthropic's Jack Clark, ahead of planned meetings in Washington; it followed similar recent proposals from Altman and Dario Amodei (A Framework for Frontier AI and the Dawning of a New Age (Hassabis, July 2026); Source: platformer.news). Hassabis's plan to lobby Washington on the proposal became public on July 16, 2026, and on July 17 Bloomberg reported that the US government was considering creating a FINRA-like watchdog to vet top AI models (Source: bloomberg.com; bloomberg.com). The essay itself, published on Hassabis's Substack, proposes that the standards body could coordinate a development slowdown "if deemed necessary" (A Framework for Frontier AI and the Dawning of a New Age (Hassabis, July 2026)). Zvi Mowshowitz responded on July 19 that "you need an SEC to your FINRA," arguing a voluntary regime that exempts internal deployment is inadequate (Source: thezvi.substack.com); critiques of the open-weight gap in self-regulation schemes are covered at AI Pre-Release Vetting. OpenAI president Greg Brockman endorsed the proposal in an interview published July 25, 2026 — "I think Demis's idea is quite good… A vacuum slows everyone down" (Source: fortune.com). Support continued to build through July 21, 2026, with endorsements from Mustafa Suleyman, Satya Nadella, and Jack Dorsey, while Dario Amodei said he favors an FAA-style federal agency instead (A Framework for Frontier AI and the Dawning of a New Age (Hassabis, July 2026); Source: axios.com). By August 28, 2026 the proposal had reached a draft executive order circulated inside the Trump administration calling for the creation of a self-regulatory organization for developers of state-of-the-art models, traced in reporting to Hassabis's framework and to a similar approach proposed by Treasury Secretary Scott Bessent; the drafting is reported to have stalled after attempts to win buy-in from senior officials and the president. The account is paywalled beyond its abstract, so the draft's terms are not established (Source: theinformation.com). See AI Pre-Release Vetting.
Hassabis attended King Charles III's meeting with AI executives at Dumfries House on September 17, 2026 (Source: reuters.com). See King Charles III.
AI for science
Hassabis distinguishes "AI for science" (AlphaFold, Isomorphic Labs) from pure language models, and has argued that the scientific use case is where AI's most durable value lies.
AGI timelines and economic transformation
Speaking at Stanford's Graduate School of Business on May 22, 2026, Hassabis said AI development stands "in the foothills of the singularity" and predicted an economic transformation 10× the scale of the Industrial Revolution, occurring 10× faster (Source: freesystems.substack.com). He has remained relatively measured on near-term labor disruption, publicly rejecting the white-collar "bloodbath" framing (see AI Labor Disruption). His economic-transformation view pairs with his decade-scale AGI timeline position.
Gary Marcus argued on June 5, 2026 that Hassabis was running two conflicting AGI timelines: "2030 plus or minus a year" at Stanford that week versus "5 to 10 years" (with a strict all-cognitive-capabilities definition) at Davos in January 2026. Marcus sided with the slower estimate and read the gap as definitional slippage, in which a looser AGI definition yields a nearer date. (Source: garymarcus.substack.com)
In a Semafor interview circulated July 13, 2026, Hassabis put AGI two to five years away, warned that biological and nuclear AI risks follow cyber risks "in the next couple of years," and called for an international body to test frontier systems before release (Source: the-ai-corner.com). See AI Pre-Release Vetting.
Relationships
- related: Google DeepMind — co-founder and CEO
- related: Gemini 3 — current flagship model under his leadership
- related: Mustafa Suleyman — DeepMind co-founder, now at Microsoft AI
- related: DeepMind Institute — director (from September 2026)
- related: Introducing the DeepMind Institute (Legg, Manyika and Hassabis, September 2026) — co-author of the Institute's founding statement, which uses the all-cognitive-capabilities definition of AGI from his July 2026 framework essay
- related: Bletchley Declaration, Seoul Commitments — signatory
- related: CAIS Statement on AI Risk — signatory
- supports: AI Safety Frameworks, Compressed 21st Century (via AI-for-science framing)
- related: AGI Timelines — decade-scale position