“If OpenAI were obliged to retrain models only on materials for which they had consent, it significantly would weaken their results, and cost a great deal of money to do that retraining.”
Our reading, the sentence that was checked:
If OpenAI had to retrain only on consented materials, its results would be significantly weakened and retraining would be very costly.
Assessment: Unresolved No deadline stated on 2026-10-29 Evidence as of
Where and when Gary Marcus wrote it
Marcus on AI, Top 5 reasons OpenAI is probably not worth 90 billion dollars, paragraph 7.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment marcus-top-5-reasons-openai-is-probably-0d189e-a2
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 2023-09-27. This page does not claim the wording was captured then.
What the claim commits to
- Kind
- Conditional forecast
- Deadline, as the speaker wrote it
- None stated. This site does not add one.
- Deadline state when assessed
- No deadline stated
- Deadline state on 2026-10-29
- No deadline stated (the release’s reference date)
Conditions
- if OpenAI were obliged to retrain only on consented materials (antecedent, stated by the speaker)
From the speaker
- If OpenAI were obliged
- retrain models only on materials for which they had consent
- significantly would weaken their results
- cost a great deal of money
Our interpretation
- The antecedent requires a binding obligation to retrain using exclusively consented materials. A lawsuit, a voluntary licence or a restriction concerning particular works is insufficient by itself.
- Weakened results means a material decline on comparable evaluations of the affected models, attributable to the consent restriction.
- The expense prediction concerns the incremental cost of the required retraining.
- Both performance decline and substantial expense must occur to support the complete forecast. No deadline or accounting basis is supplied.
Counts as supported if: After the antecedent is documented, comparable before-and-after evaluations show materially weaker results caused by the restriction, and appropriately scoped evidence establishes substantial incremental retraining costs.
Counts as contradicted if: After the antecedent and completed compliant retraining are documented, credible evaluations show no material performance decline, or appropriately scoped evidence establishes that retraining costs were not substantial.
Retrospective: these criteria were written after the outcome was public.
Current assessment
Did the forecast come true?
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 deadline stated
- Assessment recorded
- Policy version
- 2026-10-28.7
Rationale
The Munich court's November 11, 2025 summary reports infringement findings concerning particular song lyrics retained in OpenAI models and reproduced in outputs [e1]. It does not establish a binding obligation to retrain exclusively on consented materials, completed compliant retraining, or resulting performance and cost evidence [e1]. This evidence therefore leaves both the forecast's trigger and its predicted consequences unverified [e1].
Limitations
- The source is an official court summary, not the complete judgment [e1].
- The search does not comprehensively establish the status of every relevant court docket.
- The packet contains no qualifying implementation records, comparable before-and-after evaluations or incremental retraining cost records.
Evidence (1 item)
Context only: Munich Regional Court I, Bavarian Ministry of Justice,
Die streitgegenständlichen Liedtexte seien reproduzierbar in den Modellen festgelegt.
Why it matters here: The court reports findings concerning particular lyrics retained in models and relief for their infringement. Its account does not establish a requirement to retrain exclusively on consented materials, and supplies no resulting retraining costs or performance comparisons.
- Publisher
- Munich Regional Court I, Bavarian Ministry of Justice
- Kind of source
- Secondary summary
- Describes events of
- Public from
What would settle it
A documented enforceable obligation to retrain exclusively on consented materials, completed compliant retraining, comparable evaluations isolating the restriction's effect, and scoped incremental cost records. Material performance decline and substantial cost would support the forecast; no material decline or non-substantial cost would contradict it.
Where it was said (1)
Said once in the sources this register covers.
: Marcus on AI, Top 5 reasons OpenAI is probably not worth 90 billion dollars
If OpenAI were obliged to retrain models only on materials for which they had consent, it significantly would weaken their results, and cost a great deal of money to do that retraining.
- Where
- paragraph 7
- Published
- 2023-09-27
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment marcus-top-5-reasons-openai-is-probably-0d189e-a2: 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-top-5-reasons-openai-is-probably-0d189e-a2.