“ChatGPT went utterly wild when Ernie pushed on the last one, omitting some that are there and and confabulating a standalone island in the middle of the kitchen that simply isn’t there:”
Our reading, the sentence that was checked:
When questioned about the kitchen image, ChatGPT's vision model left out drawers that are there and made up a kitchen island that is not there.
Assessment: Unresolved Evidence as of
Where and when Gary Marcus wrote it
Marcus on AI, Hello, Multimodal Hallucinations, paragraph 8.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment marcus-hello-multimodal-hallucinations-ce5317-a1
We recorded 11 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 2023-10-21. 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)
Conditions
- Julia Child's kitchen image test (qualifier, stated by the speaker)
From the speaker
- when Ernie pushed on the last one
- omitting some that are there
- confabulating a standalone island
- that simply isn’t there
Our interpretation
- The claim concerns the particular recorded exchange, not ChatGPT's general reliability.
- Both components must hold: the response omitted visible drawers and asserted a nonexistent standalone island.
- The post describes the experiment as occurring 'the other day'. I use October 2023 through publication as the approximate target period; the exact test date is unspecified.
Counts as supported if: An independently accessible record of the original prompts, image and responses establishes both omitted visible drawers and an invented standalone island.
Counts as contradicted if: The original record establishes that the alleged omissions or island description did not occur, or that the described island was actually present.
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 searches located the authors' account but no independent record sufficient to verify both alleged errors. Evidence that GPT-4V hallucinated elsewhere does not establish this particular incident. This is an evidence gap, not a finding that Marcus was wrong.
Limitations
- The page reader does not reproduce the relevant image and transcript screenshots.
- A later recreation would not prove what the original exchange contained.
Evidence (0 items)
No evidence item settled this claim by the cutoff.
What would settle it
An independently accessible original conversation record, including its input image and complete follow-up prompts, permitting direct comparison with the alleged omissions and invented island.
Where it was said (1)
Said once in the sources this register covers.
: Marcus on AI, Hello, Multimodal Hallucinations
ChatGPT went utterly wild when Ernie pushed on the last one, omitting some that are there and and confabulating a standalone island in the middle of the kitchen that simply isn’t there:
- Where
- paragraph 8
- Published
- 2023-10-21
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment marcus-hello-multimodal-hallucinations-ce5317-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-hello-multimodal-hallucinations-ce5317-a1.