ClaimsDemis Hassabis

“Our paper published in @Nature today is a good example.”

Demis Hassabis, DeepMind CEO, on X 15 January 2020 Post by @demishassabis on X, 2020-01-15 (X)

Factual Science and research methods

A DeepMind paper on AI and neuroscience was published in Nature on 2020-01-15.

Resolved true The record shows it happened or was so.

Factual: Statements about the present or the past, checked against the record: did they hold up?

Said for an organisation they lead or work for; it counts on both records.

What would count as true or false

True if a DeepMind-authored paper linking AI and neuroscience appeared in Nature on 2020-01-15. The expected paper is on distributional reinforcement learning and dopamine (Dabney et al.); the date can be checked against Nature's published-online date. Mostly true if the paper appeared in Nature within a day or so of that date, or is in a Nature-branded context with a minor detail off. Mostly false if the paper was published in another journal or on a clearly different date. False if no such DeepMind AI–neuroscience paper was published in Nature around then.

Set before the research began.

Evidence (4)

  1. E1 A distributional code for value in dopamine-based reinforcement learning - Nature Nature (Springer Nature) · 15 January 2020

    “Dabney, W., Kurth-Nelson, Z., Uchida, N. et al. A distributional code for value in dopamine-based reinforcement learning. Nature 577, 671–675 (2020).”
  2. E2 A distributional code for value in dopamine-based reinforcement learning - PMC PubMed Central (U.S. National Library of Medicine, NIH) · 15 January 2020

    “Nature. 2020 Jan 15;577(7792):671–675. doi: 10.1038/s41586-019-1924-6”
  3. E3 A distributional code for value in dopamine-based reinforcement learning - PMC PubMed Central (U.S. National Library of Medicine, NIH) · 15 January 2020

    “Find articles by Demis Hassabis”
  4. E4 A distributional code for value in dopamine-based reinforcement learning - Nature Nature (Springer Nature) · 15 January 2020

    “Here we propose an account of dopamine-based reinforcement learning inspired by recent artificial intelligence research on distributional reinforcement learning4,5,6.”

Reasoning

The PubMed Central record dates the Nature paper "A distributional code for value in dopamine-based reinforcement learning" (Dabney et al., Nature 577:671–675) to 15 January 2020 (E1, E2). The abstract describes it as an account of dopamine-based reinforcement learning inspired by AI research on distributional reinforcement learning (E4), so the paper links AI and neuroscience. The PMC record lists Demis Hassabis among the authors (E3), which supports DeepMind authorship. This matches the expected paper and date in the criteria.

Cite this claim

Demis Hassabis, “A DeepMind paper on AI and neuroscience was published in Nature on 2020-01-15.” (said 15 January 2020). Verdict: resolved true. Reliability API, claim c_6d9ab9e3df7d, method 0.1.0+c876030c, dataset release v1.1.2. https://epistacklabs.com/claims/c_6d9ab9e3df7d

This claim is in dataset release v1.1.2, with the same verdict.

Reply to this claim

For Demis Hassabis, or someone speaking for them: a correction or counter-evidence. Replies are read before anything is shown; an approved reply shows your name, role, date, statement and link. Your email address is never shown and is deleted 30 days after the decision (privacy).

More factual claims by Demis Hassabis

  1. Mostly true Factual 10 April 2026
    “what’s important about them is their 3D structure.”

    A protein's 3D structure is what determines, or partially determines, its function.

  2. Resolved true Factual 10 April 2026
    “So over 3 million scientists are now using AlphaFold.”

    Over 3 million scientists are now using AlphaFold.

  3. Resolved true Factual 10 April 2026
    “Only about 10% of drugs actually get through all the clinical stages.”

    Only about 10% of drugs get through all the clinical stages.

  4. Mostly true Factual 25 November 2025
    “The AlphaFold Server is empowering non-commercial researchers globally to harness this technology, accelerating their ability to formulate and test new hypotheses.”

    The AlphaFold Server is available to non-commercial researchers globally for predicting structures and interactions.

  5. Mostly true Factual 25 November 2025
    “Our AlphaProteo model can design novel, high-strength protein binders that target diverse molecules – including those associated with cancer and diabetes.”

    AlphaProteo can design novel, high-strength protein binders targeting diverse molecules, including ones associated with cancer and diabetes.

All of them

Verdicts are made automatically; how verdicts are made.