“OpenAI has never disclosed those numbers for GPT-4; I can scarcely imagine what they will be for GPT-5.”
Our reading, the sentence that was checked:
OpenAI has never disclosed the water consumption figures for GPT-4.
Assessment: Supported Evidence as of
Where and when Gary Marcus wrote it
Marcus on AI, Seven reasons why the world should say No to Sam Altman, paragraph 11.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment marcus-seven-reasons-why-the-world-should-023ed2-a1
We recorded 15 claims from this post. See every claim from this post.
The wording above is the version this site observed on 2026-10-29. The source may have been edited after 2024-02-10. This page does not claim the wording was captured then.
What the claim commits to
- Kind
- Observation
- Deadline, as the speaker wrote it
- None stated. This site does not add one.
- Deadline state when assessed
- No horizon
- Deadline state on 2026-10-29
- No horizon (the release’s reference date)
From the speaker
- OpenAI has never disclosed
- those numbers for GPT-4
Our interpretation
- Those numbers refers principally to GPT-4 training water consumption, following the comparison with estimated GPT-3 training water use.
- Disclosed means made public by OpenAI. Third-party estimates or Microsoft-wide water totals do not constitute an OpenAI disclosure of GPT-4's consumption.
- Never describes the public record through February 10, 2024, including any pre-release disclosures; no artificial starting date is imposed.
- The remark about scarcely imagining GPT-5 figures supplies no separate testable forecast.
Counts as supported if: Review of relevant OpenAI disclosures, corroborated by contemporary research on reporting gaps, finds no publicly disclosed GPT-4 training-water figure through the post date.
Counts as contradicted if: An authenticated OpenAI disclosure published by February 10, 2024 gives a quantitative GPT-4 training-water figure.
Retrospective: these criteria were written after the outcome was public.
Current assessment
Was it true for the period described?
Outcome: Supported
The dated evidence matches what the claim committed to, under the criteria frozen before grading.
Scored on a plain reading. The speaker named no measure for this claim, so it was judged on the reading an informed reader would take, stated below under our interpretation. Both reviews had to find that reading fair.
- Evidence as of
- Deadline state when assessed
- No horizon
- Assessment recorded
- Policy version
- 2026-10-28.7
Rationale
The public-record review supports the narrow disclosure claim. OpenAI's release materials contain no GPT-4 training-water figure, and contemporary independent assessments corroborate the reporting gap. No qualifying disclosure dated by February 10, 2024 was located. Microsoft-wide totals and outside estimates do not contradict a claim about what OpenAI disclosed.
Limitations
- A negative disclosure claim cannot exclude an obscure or subsequently removed publication with absolute certainty.
- The independent assessments precede the post; searches for intervening disclosures found none.
- This finding does not establish how much water GPT-4 actually consumed.
Evidence (3 items)
Supports the claim: Stanford CRFM and collaborating researchers, via arXiv,
no company scores points
Why it matters here: The assessment includes OpenAI's GPT-4. The following sentence identifies broader environmental impact among the universally undisclosed indicators, and the preceding discussion explicitly includes water use.
- Publisher
- Stanford CRFM and collaborating researchers, via arXiv
- Kind of source
- Preprint, not peer reviewed
- Describes events of
- Public from
Supports the claim: Pengfei Li, Jianyi Yang, Mohammad A. Islam and Shaolei Ren, via arXiv,
limited public data
Why it matters here: Researchers studying model water consumption used GPT-3 because GPT-4's public data were limited. This corroborates the disclosure gap, although it does not alone prove the universal negative.
- Publisher
- Pengfei Li, Jianyi Yang, Mohammad A. Islam and Shaolei Ren, via arXiv
- Kind of source
- Preprint, not peer reviewed
- Describes events of
- Public from
Context only: OpenAI,
capabilities, limitations, and safety properties
Why it matters here: The report withholds detailed information about training resources. Checking its water references located no training-water figure; the release page likewise yielded no water disclosure.
- Publisher
- OpenAI
- Kind of source
- Company disclosure: the company describing itself, not an independent finding
- Describes events of
- Public from
Where it was said (1)
Said once in the sources this register covers.
: Marcus on AI, Seven reasons why the world should say No to Sam Altman
OpenAI has never disclosed those numbers for GPT-4; I can scarcely imagine what they will be for GPT-5.
- Where
- paragraph 11
- Published
- 2024-02-10
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment marcus-seven-reasons-why-the-world-should-023ed2-a1: Current. Outcome: Supported. Evidence as of 2026-10-28. Public date not recorded.
Recorded events
Assessment publishedBy this site
First assessment published after both reviews and the final-text check passed.
About assessment marcus-seven-reasons-why-the-world-should-023ed2-a1.