“The first example illustrates that ChatGPT with DALL-E 3 is unable to reliably associate labels correctly with the parts of an object.”
Our reading, the sentence that was checked:
ChatGPT with DALL-E 3 cannot reliably associate labels correctly with the parts of an object.
Assessment: Unresolved Evidence as of
Where and when Gary Marcus wrote it
Marcus on AI, Don’t Ride This Bike! Generative AI’s persistent trouble with compositionality and parts, paragraph 28.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment marcus-dont-ride-this-bike-generative-ais-5742c2-a1
We recorded 10 claims from this post. See every claim from this post.
The wording above is the version this site observed on 2026-10-28. The source may have been edited after 2024-12-08. 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
- based on a small number of examples (qualifier, stated by the speaker)
From the speaker
- ChatGPT with DALL-E 3
- unable to reliably associate labels correctly with the parts of an object
- on the basis of a small number of examples
Our interpretation
- The plain reading is recurring failure to attach labels to the correct object parts, not an inability to produce any correct diagram.
- The post dates the underlying experiments to November 10, 2024 and describes the system's performance in November 2024. November 2024 is therefore the target period.
- Label placement and identification must be assessed separately from spelling and image quality. No numerical reliability threshold is supplied.
Counts as supported if: Dated records of the November 2024 system show repeated incorrect associations between labels and object parts across relevant attempts or tasks.
Counts as contradicted if: Comparable evaluations of that system show consistently correct label-to-part associations, with the reported failures established as exceptional or incorrectly scored.
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 independent report documents related diagram problems, but it tests an earlier Bing deployment and does not isolate label-to-part association. The November bicycle transcript is the underlying record for the post itself, and its images were not independently verified here. These records do not securely establish or refute the claimed reliability of November ChatGPT with DALL-E 3.
Limitations
- Unresolved does not mean the bicycle example was correct.
- This assessment cannot estimate a failure rate.
Evidence (1 item)
Context only: Public Citizen,
generating nonsense words.
Why it matters here: The report documents problems in labeled diagrams generated with Bing's DALL-E 3. Its earlier deployment and emphasis on nonsensical lettering do not settle November ChatGPT performance on associating readable labels with parts.
- Publisher
- Public Citizen
- Kind of source
- Demonstration
- Describes events of
- Public from
What would settle it
Verification of the original November diagrams and prompts, or a comparable dated evaluation that scores repeated label-to-part associations.
Where it was said (1)
Said once in the sources this register covers.
: Marcus on AI, Don’t Ride This Bike! Generative AI’s persistent trouble with compositionality and parts
The first example illustrates that ChatGPT with DALL-E 3 is unable to reliably associate labels correctly with the parts of an object.
- Where
- paragraph 28
- Published
- 2024-12-08
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment marcus-dont-ride-this-bike-generative-ais-5742c2-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-dont-ride-this-bike-generative-ais-5742c2-a1.