“That has meant, for example, no tangible progress on getting the big companies to commit to revealing what data their models were trained on; we have transparency in name, but not in deed.”
Our reading, the sentence that was checked:
There has been no tangible progress in getting big AI companies to commit to disclosing their training data.
Assessment: Unresolved Evidence as of
Where and when Gary Marcus wrote it
Marcus on AI, Could we all be doomed, without ranked-choice voting?, paragraph 4.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment marcus-could-we-all-be-doomed-without-ranked-3eea03-a1
We recorded 9 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 2023-08-29. 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
- no tangible progress
- getting the big companies to commit
- revealing what data their models were trained on
- transparency in name, but not in deed
Our interpretation
- The ordinary reading concerns concrete progress toward meaningful identification of training-data sources by major AI companies.
- Meaningful disclosure need not publish every training record, but must identify sources more specifically than an assertion that a model used public or licensed data.
- The context concerns US governance efforts. Voluntary company commitments can nevertheless count.
- An existing research disclosure is relevant counterevidence to the broader transparency language, but is not automatically evidence that policymakers secured a continuing disclosure commitment.
- The claim describes cumulative progress through 2023-08-29, not a reporting quarter. Its wording supplies no starting date.
Counts as supported if: The principal relevant commitments omitted meaningful training-data disclosure, with sufficiently broad corroborating records to establish that no concrete disclosure commitment had been secured from major companies.
Counts as contradicted if: A concrete agreement, adopted requirement, or public major-company commitment by the post required meaningful identification of model training-data sources.
Retrospective: these criteria were written after the outcome was public.
Current assessment
Was it true for the period described?
Outcome: Unresolved
The record was checked, and the evidence available by the cutoff does not settle it.
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 White House commitments omitted training-data disclosure, which supports Marcus's criticism of that initiative. However, Meta had already published meaningful source information for LLaMA. Existing disclosure qualifies the broader claim of transparency only in name, but does not itself prove that a disclosure commitment had been secured. The records therefore do not establish either the claimed absence of progress across major companies or a decisive counterexample to the narrower commitment claim.
Limitations
- The result depends on distinguishing existing disclosures from commitments to disclose.
- Neither generic source categories nor a full public training corpus is imposed as the sole test.
- Later transparency requirements cannot settle this historical observation.
Evidence (2 items)
Supports the claim: The White House,
capabilities, limitations
Why it matters here: The public-reporting commitment covers capabilities, limitations, and uses. The complete commitment list omits training-data disclosure, supporting the claim for this central government initiative but not proving absence across every company or initiative.
- Publisher
- The White House
- Kind of source
- News report
- Describes events of
- Public from
Context only: Meta AI authors, via arXiv,
We use the public GitHub dataset available on Google BigQuery.
Why it matters here: The paper identifies concrete training sources and describes their proportions and processing. That is substantive existing transparency, but it does not establish that government secured a continuing disclosure commitment.
- Publisher
- Meta AI authors, via arXiv
- Kind of source
- Company disclosure: the company describing itself, not an independent finding
- Describes events of
- Public from
What would settle it
A dated agreement or company policy establishing a meaningful disclosure commitment by August 29, 2023, or a sufficiently complete contemporaneous review of the relevant negotiations and company commitments showing none had been secured.
Where it was said (1)
Said once in the sources this register covers.
: Marcus on AI, Could we all be doomed, without ranked-choice voting?
That has meant, for example, no tangible progress on getting the big companies to commit to revealing what data their models were trained on; we have transparency in name, but not in deed.
- Where
- paragraph 4
- Published
- 2023-08-29
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment marcus-could-we-all-be-doomed-without-ranked-3eea03-a1: Current. Outcome: Unresolved. 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-could-we-all-be-doomed-without-ranked-3eea03-a1.