A federal judge has accepted 56 amicus briefs in the copyright case over OpenAI’s AI training, giving a wide range of organisations and individuals a chance to argue how copyright law should treat the use of protected works. Their arguments show why this case could matter well beyond one company, though they do not decide its outcome.
OpenAI Watch analysis
What happened
Judge Sidney Stein accepted the briefs on 6 November in the federal litigation over OpenAI’s use of copyrighted material to train AI models, according to Tech Policy Press. Some briefs support the authors and publishers bringing claims; others support OpenAI and Microsoft. Several declare no side, or say they support neither party.
Tech Policy Press has compiled a directory of all 56 briefs, with declared positions, selected quotations and data on the court precedents they cite. Its analysis finds that citation patterns cut across the dispute, but a citation is not necessarily an endorsement: a brief may cite a ruling to argue that it does not apply. Explore the briefs and analysis.
Why it matters
The case asks whether training AI models on copyrighted material can qualify as fair use. OpenAI and Microsoft argue that models learn from works rather than reproduce them; authors and publishers argue that training involves unlicensed copying. The accepted briefs add perspectives from groups including libraries, technology trade bodies, creative unions and AI companies. They are arguments for the court to consider, not findings about what happened or a preview of the ruling.
The analysis itself comes with an unusually useful warning label. Tech Policy Press says it used Claude Opus 5.5 to extract citations and make individual decisions about ambiguous references. Those decisions were not reviewed by an editor, although an editor checked a sample of quotations against the briefs. The underlying data and methods are available for readers who want to inspect the work rather than take a chart’s word for it.
Our read
This is a valuable map of the arguments, not a scorecard predicting the winner. The important distinction is between what the briefs claim and what the judge ultimately accepts. If you are following the case, the directory and downloadable data offer a better starting point than treating a citation count as a legal verdict. Even a tidy visualisation cannot make a contested point of law tidy.
What to watch
- Whether the court’s decision on the fair-use question addresses the arguments raised in the briefs.
- Replies to the briefs, due on 20 November, according to Tech Policy Press.
- The court’s treatment of claims about training, copying and the purpose of AI models.
Discussion spark: Should AI training on copyrighted work count as fair use when the resulting models do not reproduce that work, or should training require permission from rights holders?
Sources and evidence
- Friends of the Court Flock to the OpenAI Copyright Fight – Tech Policy Press (12 November 2026, 00:24 UTC)
OpenAI Watch is independently operated by WittyWires. It is not affiliated with, endorsed by, or operated by OpenAI.