“no robust solution to hallucinations”
Our reading, the sentence that was checked:
There is no robust solution to LLM hallucinations.
Assessment: Supported Evidence as of
Where and when Gary Marcus wrote it
Marcus on AI, Could China devastate the US without firing a shot?, paragraph 2.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment marcus-could-china-devastate-the-us-without-e898be-a1
We recorded 9 claims from this post. See every claim from this post.
The wording above is the version this site observed on 2026-10-28. The source may have been edited after 2025-10-27. 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 robust solution to hallucinations
Our interpretation
- Plain reading: as of the post date (October 2025), no method reliably eliminated LLM hallucinations; leading models still produced confident false statements.
- Target period: the state just before the post, September to October 2025, the most recent period public evidence covered. This is an ongoing-state claim with no named period.
Counts as supported if: Leading developers or independent evaluations report that current frontier models still hallucinate and that the problem remains unsolved.
Counts as contradicted if: A credible source shows a method that reliably eliminates hallucinations across general use for leading models.
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
In September 2025, about seven weeks before the post, OpenAI wrote that GPT-5 has significantly fewer hallucinations, 'but they still occur'. It added that 'Hallucinations remain a fundamental challenge for all large language models.' It also called the problem 'stubbornly hard to fully solve.' A leading developer's own statement that the problem persists in its newest model supports the claim that there was no robust solution at the time. Rates have fallen, but the claim is about a solution, not progress.
Limitations
- 'Robust' has no threshold; partial mitigations exist and rates are falling.
- Rests on one developer's statement, though it covers all LLMs.
Evidence (2 items)
Supports the claim: OpenAI,
GPT‑5 has significantly fewer hallucinations especially when reasoning, but they still occur. Hallucinations remain a fundamental challenge for all large language models, but we are working hard to further reduce them.
Why it matters here: OpenAI itself says hallucinations still occur in GPT-5 and remain a fundamental challenge for all LLMs. That supports 'no robust solution'.
- Publisher
- OpenAI
- Kind of source
- Company disclosure: the company describing itself, not an independent finding
- Describes events of
- Public from
Supports the claim: OpenAI,
Even as language models become more capable, one challenge remains stubbornly hard to fully solve: hallucinations.
Why it matters here: The developer describes hallucinations as stubbornly hard to fully solve. That supports the claim.
- 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, Could China devastate the US without firing a shot?
no robust solution to hallucinations
- Where
- paragraph 2
- Published
- 2025-10-27
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment marcus-could-china-devastate-the-us-without-e898be-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-could-china-devastate-the-us-without-e898be-a1.