“None of the top labs actually is remotely transparent, governance basically doesn’t exist yet, and there are basically no safeguards actually in place.”
Our reading, the sentence that was checked:
None of the top AI labs is remotely transparent.
Assessment: Contradicted Evidence as of
Where and when Gary Marcus wrote it
Marcus on AI, The Open Letter Controversy, paragraph 25.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment marcus-the-open-letter-controversy-b386ab-a1
We recorded 12 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 2023-03-30. 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
- None of the top labs
- actually is remotely transparent
Our interpretation
- The ordinary reading is that no leading AI laboratory provided meaningful public information permitting scrutiny of its models' development, capabilities, or risks.
- Leading laboratories include Meta, OpenAI, Google/DeepMind, and Anthropic. Marcus supplies neither a list nor a transparency threshold.
- Substantive disclosures across training data, methods, and evaluation count as meaningful transparency; complete disclosure of every institutional activity is not required.
- For this ongoing observation, the evidence window is the current calendar quarter through publication, January 1–March 30, 2023. This approximates the contemporary state, rather than a quarterly average.
Counts as supported if: Contemporary records show that leading laboratories uniformly withheld substantive information needed to scrutinize model development and evaluation, with public disclosures largely superficial.
Counts as contradicted if: At least one clearly leading laboratory published substantive information across several relevant dimensions, allowing meaningful scrutiny beyond promotional claims.
Retrospective: these criteria were written after the outcome was public.
Current assessment
Was it true for the period described?
Outcome: Contradicted
The dated evidence goes against 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
Meta's contemporary LLaMA paper disclosed substantive information about data, training, compute, evaluations, and risks. Under the ordinary reading of meaningful model transparency, that defeats 'none'. OpenAI's explicit omissions show that serious transparency gaps existed. The judgment concerns the universal assertion, not whether disclosure was adequate overall.
Limitations
- The result depends on counting substantive model disclosures as meaningful laboratory transparency.
- The evidence does not establish complete transparency at Meta or any other laboratory.
Evidence (2 items)
Goes against the claim: Meta AI / arXiv,
mixture of several sources
Why it matters here: Meta disclosed named data sources and their proportions, preprocessing, architecture, training settings, hardware, and benchmark results, including bias and toxicity assessments. These disclosures provide a substantive counterexample to the assertion that no leading laboratory offered meaningful transparency.
- Publisher
- Meta AI / arXiv
- Kind of source
- Preprint, not peer reviewed
- Describes events of
- Public from
Context only: OpenAI / arXiv,
capabilities, limitations, and safety properties
Why it matters here: The surrounding passage expressly withholds architecture details, model size, hardware, training compute, dataset construction, and training methods. This supports Marcus's concern about important disclosure gaps, but does not establish his universal claim about every leading laboratory.
- Publisher
- OpenAI / arXiv
- Kind of source
- Preprint, not peer reviewed
- Describes events of
- Public from
Where it was said (1)
Said once in the sources this register covers.
: Marcus on AI, The Open Letter Controversy
None of the top labs actually is remotely transparent, governance basically doesn’t exist yet, and there are basically no safeguards actually in place.
- Where
- paragraph 25
- Published
- 2023-03-30
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment marcus-the-open-letter-controversy-b386ab-a1: Current. Outcome: Contradicted. 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-the-open-letter-controversy-b386ab-a1.