Scores by kind of claim
Each kind of claim is scored on its own, out of 100: the share that held up, with the range it likely falls in. Fewer resolved claims give a wider range.
-
Factual claims
Statements about the present or the past, checked against the record: did they hold up?
81 76 87 -
Predictions
Statements about events outside the person's control, checked once they fell due: did they come true?
Insufficient record: 2 predictions resolved, 5 needed for a score.
-
Promises
Commitments to the person's or organisation's own future action, with something to deliver: was it delivered?
65 45 83
All types combined: 80 (likely 75 to 85, high confidence), from 84 claims: 78 factual claims, 6 promises. It mixes kinds of claim in different proportions for different people, so compare people kind by kind.
Demis Hassabis's claims
5 claims
| Verdict | Kind | Said | Claim | Topic |
|---|---|---|---|---|
| Mixed | Factual | 10 Apr 2026 | Hassabis got into AI originally because he was interested in big questions such as the nature of reality and consciousness. Factual · 10 Apr 2026 | Their own work or organisation |
| Mixed | Factual | 25 Nov 2025 | Protein misfolding can cause loss of function and lead to diseases such as Alzheimer's and Parkinson's. Factual · 25 Nov 2025 | Biosecurity |
| Mixed | Factual | 3 Aug 2025 | The success of the AI field in the last few years has attracted more attention, resources and talent. Factual · 3 Aug 2025 | AI, the economy and work |
| Mixed | Factual | Dec 2024 | Hassabis's team has practised open science and open access with all of its scientific work. Factual · Dec 2024 | Their own work or organisation |
| Mixed | Factual | 28 Feb 2024 | Large models of the Gemini and GPT-4 generation generalize to new constructs rather than merely memorizing. Factual · 28 Feb 2024 | AI capabilities |
“We’re huge believers in open science and open access, and we’ve done that with all of our scientific work.”
Demis Hassabis, CEO of Google DeepMind, interviewee December 2024 Transcript from an interview with Demis Hassabis (NobelPrize.org)
Hassabis's team has practised open science and open access with all of its scientific work.
Mixed Partly true and partly false in roughly equal measure.
Said for an organisation they lead or work for; it counts on both records.
What would count as true or false
Test: as of December 2024, had Google DeepMind (Hassabis's team, including the earlier DeepMind) made all of its scientific work available openly, meaning research published openly or open-access and key scientific outputs released for others to use? Judge the claim as stated, including the word 'all', but limited to scientific work, not commercial products such as Gemini. True: the evidence shows the team's scientific work (for example AlphaFold, AlphaFold DB, AlphaFold 3, GNoME, AlphaMissense, weather models) was consistently published openly, with code, models or data made available, and no notable scientific output was kept closed. Mostly true: open science was the clear and consistent practice, with only minor exceptions such as a delayed release or a narrow restriction. Mixed: a substantial share of the scientific work was open, but significant outputs were withheld or restricted. For example, AlphaFold 3 was first released without code and later with weights for academic use only, and some papers were published without code. Taken together, 'all' clearly overstates the practice, though it rests on real open releases. Mostly false: significant scientific work was kept closed and the openness was selective, so the claim overstates a partial practice. False: the evidence shows the team's scientific work was generally not published openly or not made accessible.
Evidence
- Daily briefing: AlphaFold3 is now open source Nature (Nature Briefing, Springer Nature) · 12 November 2024
“The initial publication, six months ago, of AlphaFold3 without its code drew criticisms from scientists, who said the move undermined reproducibility.”
- GitHub - google-deepmind/materials_discovery GitHub (repository owned by the google-deepmind organisation) · 28 November 2023 Own record
“Data contained in the Graph Networks for Materials Exploration Database is available for use under the terms of the Creative Commons Attribution Noncommercial 4.0 International Licence (CC BY NC 4.0).”
- AlphaFold Wins Nobel Prize, Gains Functionality, Drops Open Access Engineering (Chinese Academy of Engineering, journal) · 15 February 2025
“AlphaFold Wins Nobel Prize, Gains Functionality, Drops Open Access”
- Report: Google’s DeepMind is withholding AI research for competitive advantage Sherwood News · 1 April 2025
“the organization is breaking with that tradition of scientific sharing by withholding AI research for Google’s strategic advantage in a very competitive market.”
Reasoning
E1 (Nature) reports that AlphaFold3 was first published without its code, which scientists criticised as undermining reproducibility, and that it was made open source only later, in November 2024. E3, a journal piece dated February 2025 (after the claim) and known only by its headline, describes AlphaFold as one that "Drops Open Access". E2 shows the GNoME materials data was released, but only under a non-commercial licence. Taken together, the evidence shows real open releases but also significant outputs that were withheld or restricted, so the word 'all' overstates the practice; this matches the 'mixed' standard in the criteria.
Cite this claim
Demis Hassabis, “Hassabis's team has practised open science and open access with all of its scientific work.” (said December 2024). Verdict: mixed. Reliability API, claim c_1c3ca94ea27c, method 0.1.0+c876030c. https://epistacklabs.com/claims/c_1c3ca94ea27c
Claims without a verdict
Is a verdict here wrong, or do you speak for Demis Hassabis? Reply on the claim's own page, with the form at its foot (how replies work). Verdicts are made automatically; how verdicts and scores are made.