“For another, although OpenAI may be tracking errors like these for internal purposes, the scientific community has no access to whatever they have collected, which leaves it in the lurch.”
Our reading, the sentence that was checked:
The scientific community has no access to whatever error data OpenAI collects on ChatGPT.
Assessment: Unresolved Evidence as of
Where and when Gary Marcus wrote it
Marcus on AI, Large Language Models like ChatGPT say The Darnedest Things, paragraph 11.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment marcus-large-language-models-like-chatgpt-3df6dc-a1
We recorded 10 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 2023-01-09. 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 may be tracking errors like these for internal purposes
- the scientific community has no access
- whatever they have collected
Our interpretation
- The plain reading concerns outside researchers' access to OpenAI's collected ChatGPT error records, rather than access to the chatbot or independently collected examples.
- The possible existence of internal tracking is hedged; the asserted absence of research access is the factual proposition being assessed.
- For this ongoing access state, use the January 9, 2023 snapshot immediately before publication.
- The statement about leaving science in the lurch is evaluative and is not separately scored.
Counts as supported if: Dated access policies, company statements, or well-documented research-access records establish that OpenAI's collected ChatGPT error data were unavailable to outside researchers at the post date.
Counts as contradicted if: A documented release or research-access arrangement provided outside scientists access to that ChatGPT error collection, or a substantive part of it, by the post date.
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
OpenAI's contemporary rules document a feedback collection process and internal judging. They do not state whether outside scientists could obtain the collected records. No matching public release was found, but that absence cannot establish the broader assertion that the scientific community had no access.
Limitations
- Earlier OpenAI feedback datasets do not answer this ChatGPT-specific claim.
- Public chatbot access is not access to OpenAI's internal collection.
- Restricted research access may leave little public documentation.
Evidence (1 item)
Context only: OpenAI,
will review the Submissions.
Why it matters here: The launch-linked rules describe submission of problematic outputs and review by OpenAI employees. Internal review does not establish that external researchers had no access through another arrangement.
- Publisher
- OpenAI
- Kind of source
- Company disclosure: the company describing itself, not an independent finding
- Describes events of
- Public from
What would settle it
A dated OpenAI access policy or statement covering the collection, documented research-access decisions, or a verifiable release or access agreement in force by January 9, 2023.
Where it was said (1)
Said once in the sources this register covers.
: Marcus on AI, Large Language Models like ChatGPT say The Darnedest Things
For another, although OpenAI may be tracking errors like these for internal purposes, the scientific community has no access to whatever they have collected, which leaves it in the lurch.
- Where
- paragraph 11
- Published
- 2023-01-09
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment marcus-large-language-models-like-chatgpt-3df6dc-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-large-language-models-like-chatgpt-3df6dc-a1.