“OpenAI's ChatGPT and Anthropic's Claude are dependent on a constant flow of training data to improve their models, to the point that it's effectively impossible for them to operate without violating copyright.”
Our reading, the sentence that was checked:
OpenAI's ChatGPT and Anthropic's Claude depend on a constant flow of training data to improve their models, to the point that it is effectively impossible for them to operate without violating copyright.
Assessment: Unresolved Evidence as of
Where and when Ed Zitron wrote it
Where's Your Ed At, Are We Watching The Internet Die?, paragraph 34.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment wyea-are-we-watching-the-internet-die-ba4560-a1
We recorded 29 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 2024-03-11. This page does not claim the wording was captured then.
What the claim commits to
- Kind
- Causal claim
- 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
- effectively (qualifier, stated by the speaker)
From the speaker
- OpenAI's ChatGPT and Anthropic's Claude
- dependent on a constant flow of training data to improve their models
- effectively impossible for them to operate without violating copyright
Our interpretation
- The claim concerns the practical requirements of developing and improving these services as of March 11, 2024.
- In context, operating includes maintaining competitive models through further training. It does not mean that every individual answer requires newly collected training data.
- Violating copyright means infringement, not merely using material protected by copyright.
- Practical necessity is stronger than evidence that infringement occurred: it requires showing that workable lawful alternatives were unavailable.
Counts as supported if: Evidence establishes that continued improvement of ChatGPT and Claude practically required fresh training material that could not lawfully be obtained or used, with applicable legal findings supporting the infringement premise.
Counts as contradicted if: A documented, practically adequate lawful training and improvement route for the named services defeats the asserted necessity. Evidence of lawful use of only a small part of their training material is insufficient by itself.
Retrospective: these criteria were written after the outcome was public.
Current assessment
Does the evidence support the cause?
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 evidence does not establish the claimed necessity. OpenAI had an authorized training-content partnership before the post. A later court ruling treated Anthropic's training use as fair use while distinguishing pirated library copies. Later research also demonstrates useful models trained on openly licensed text. These are substantial qualifications, but none establishes a fully lawful, practically adequate replacement for both named services as they existed in March 2024. Nor does the record show that infringement was unavoidable.
Limitations
- The court ruling concerns particular works and uses under U.S. law.
- Later research on other models is context, not a direct test of the named services in the relevant period.
- A narrower reading limited to running already-trained models would require a different assessment.
Evidence (3 items)
Context only: U.S. District Court, Northern District of California; copy hosted by Hunton,
the training use was a fair use.
Why it matters here: The ruling rejects an automatic equation between copyrighted training material and infringement, while leaving pirated library copies for trial. It does not establish that every input or operation of Claude was lawful, or resolve the practical necessity asserted for both services in March 2024.
- Publisher
- U.S. District Court, Northern District of California; copy hosted by Hunton
- Kind of source
- Regulatory filing
- Describes events of
- Public from
Context only: OpenAI,
training
Why it matters here: An authorized route to additional training content existed before the post. The announcement does not show that this arrangement could replace all disputed data or sustain the entire service.
- Publisher
- OpenAI
- Kind of source
- Company disclosure: the company describing itself, not an independent finding
- Describes events of
- Public from
Context only: Kandpal et al., arXiv,
competitive performance
Why it matters here: Common Pile reports competitive smaller models trained on openly licensed text. This challenges a general technical-necessity argument, but concerns different models and a later period.
- Where
- Abstract
- Publisher
- Kandpal et al., arXiv
- Kind of source
- Preprint, not peer reviewed
- Describes events of
- Public from
What would settle it
A rights-audited training and improvement record for the named services, with evidence that lawful inputs were sufficient or that no practical lawful route existed under the relevant law.
Where it was said (1)
Said once in the sources this register covers.
: Where's Your Ed At, Are We Watching The Internet Die?
OpenAI's ChatGPT and Anthropic's Claude are dependent on a constant flow of training data to improve their models, to the point that it's effectively impossible for them to operate without violating copyright.
- Where
- paragraph 34
- Published
- 2024-03-11
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment wyea-are-we-watching-the-internet-die-ba4560-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 wyea-are-we-watching-the-internet-die-ba4560-a1.