“And the apparent copyright infringements that Reid Southen and I discussed in early January continue apace two months later, as I found in a quick experiment this morning:”
Our reading, the sentence that was checked:
The apparent copyright infringements that Marcus and Reid Southen discussed in early January 2024 were continuing apace two months later, based on a quick experiment Marcus ran on the morning of the post.
Assessment: Supported Evidence as of
Where and when Gary Marcus wrote it
Marcus on AI, OpenAI’s “Own Goal”, paragraph 25.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment marcus-openais-own-goal-03baf3-a1
We recorded 7 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 2024-03-06. 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
- apparent (qualifier, stated by the speaker)
- as found in a quick experiment this morning (qualifier, stated by the speaker)
From the speaker
- apparent copyright infringements
- discussed in early January
- continue apace two months later
- a quick experiment this morning
Our interpretation
- The plain reading is that OpenAI-powered image generators continued readily producing recognizable reproductions of protected characters or imagery in early March 2024.
- Apparent infringement means outputs visibly resembling protected material; it does not assert a court finding of infringement.
- I interpret 'continue apace' as continued ready production, rather than a measured January-to-March rate comparison. Marcus supplies no frequency measure.
- The target period is March 1–6, 2024: early March through publication, reflecting 'two months later' and 'this morning'. January tests provide comparison context.
- The scope includes Microsoft's implementation of OpenAI's image model. Independent tests can establish the broader observation without verifying Marcus's particular morning experiment.
Counts as supported if: Independent contemporaneous tests show OpenAI-powered image generation still producing recognizable protected characters or imagery during the target period, comparable in kind to the January examples.
Counts as contradicted if: Reliable contemporaneous evidence establishes that comparable reproduction had ceased during the target period, or that the alleged March examples were misidentified or came from an unrelated system. Individual refusals alone would not establish cessation.
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
Independent January testing documented recognizable character reproduction. On March 6, The Verge reported reproducing Elsa images with Microsoft's OpenAI-powered generator. That supports the ordinary reading that apparent reproductions continued into early March. Support is limited to continued production: the evidence does not measure an unchanged pace, verify Marcus's particular experiment, or establish legal infringement. January reports of stronger ChatGPT safeguards show that behavior varied by interface, but do not negate the successful March reproduction.
Limitations
- The result depends on reading 'apace' as continued ready production rather than a quantitative rate claim.
- The decisive evidence concerns Microsoft's implementation of DALL-E 3.
- The March test is dated by its contemporaneous report, without a separate execution timestamp.
- Recognizable character reproduction alone does not settle legal liability.
Evidence (2 items)
Supports the claim: The Verge,
The Verge was able to generate similar images using the tool.
Why it matters here: The article reports its own reproduction of the described Elsa images with Copilot Designer and identifies DALL-E 3 as the underlying model. This is firsthand corroboration published on March 6; the article does not give a separate test timestamp.
- Publisher
- The Verge
- Kind of source
- Independent measurement
- Describes events of
- Public from
Context only: Tom's Hardware,
Mario and Sonic
Why it matters here: The author's independent testing establishes a January comparison: an OpenAI-powered generator produced recognizable characters from descriptive prompts. The same report describes stronger refusals in ChatGPT, limiting any inference that all interfaces or prompts behaved alike.
- Publisher
- Tom's Hardware
- Kind of source
- Independent measurement
- Describes events of
- Public from
Where it was said (1)
Said once in the sources this register covers.
: Marcus on AI, OpenAI’s “Own Goal”
And the apparent copyright infringements that Reid Southen and I discussed in early January continue apace two months later, as I found in a quick experiment this morning:
- Where
- paragraph 25
- Published
- 2024-03-06
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment marcus-openais-own-goal-03baf3-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.