“Rumors that LLMs have solved then hallucination problem are greatly exaggerated.”
Our reading, the sentence that was checked:
LLMs have not solved the hallucination problem.
Assessment: Supported Evidence as of
Where and when Gary Marcus wrote it
Marcus on AI, Seven lies in four sentences, paragraph 3.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment marcus-seven-lies-in-four-sentences-99c9db-a1
We recorded 6 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 2023-10-04. 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
- LLMs
- have solved then hallucination problem
- greatly exaggerated
Our interpretation
- The ordinary reading is that factual fabrication remained an unresolved limitation of contemporary LLMs. This does not claim that every model or answer hallucinates, or that improvements had stopped.
- The context illustrates the present with examples from last week and last night. I use the last complete calendar week before publication through the post date: September 25–October 4, 2023.
- A contemporary research assessment or deployment disclosure identifying continuing factual fabrication can establish that the problem remained unsolved. It need not measure the frequency of all user encounters.
Counts as supported if: Contemporary primary research or deployment evidence identifies factual fabrication as a continuing material limitation of leading LLMs.
Counts as contradicted if: Contemporary evidence establishes that factual fabrication had effectively been eliminated across ordinary factual tasks in leading LLMs, with the alleged continuing failures disproved or confined to superseded systems.
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
Primary research revised on September 25 explicitly identifies hallucination as unresolved. OpenAI's deployment disclosure also reports factual fabrication in GPT-4V. Together they support the ordinary claim that the problem persisted near the post date. Neither source requires the stronger conclusion that most answers were false or that no progress had occurred.
Limitations
- The research was a preprint at publication.
- The deployment evidence is company-reported and includes image-based tasks.
- This assessment does not establish how common hallucinations were across all user requests.
Evidence (2 items)
Supports the claim: arXiv; Shehzaad Dhuliawala and colleagues,
an unsolved issue
Why it matters here: The revised paper identifies hallucination as an unresolved contemporary problem and reports experiments reducing it through verification. The event date identifies this current-state assessment, not the undisclosed dates of individual experimental runs.
- Publisher
- arXiv; Shehzaad Dhuliawala and colleagues
- Kind of source
- Preprint, not peer reviewed
- Describes events of
- Public from
Supports the claim: OpenAI,
prone to hallucinations
Why it matters here: OpenAI reports factual fabrication as a limitation of the model being deployed in late September. This dates the disclosed deployment limitation, not the earlier red-team sessions.
- 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 lies in four sentences
Rumors that LLMs have solved then hallucination problem are greatly exaggerated.
- Where
- paragraph 3
- Published
- 2023-10-04
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment marcus-seven-lies-in-four-sentences-99c9db-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-lies-in-four-sentences-99c9db-a1.