A federal judge in New York refused to toss authors’ output-based copyright claims against OpenAI, citing examples tied to George R. R. Martin’s works and sending the case forward.
📌 Key Takeaways
- The court denied OpenAI’s bid to dismiss output-based infringement claims.
- The order says a reasonable jury could find ChatGPT outputs substantially similar to protected books.
- The ruling highlights a ChatGPT sequel outline titled “A Dance with Shadows.”
- Fair-use questions remain undecided at this stage.
- The case is part of an MDL in Manhattan that consolidates authors’ suits.
Why The Judge Said A Jury Could See Infringement
U.S. District Judge Sidney H. Stein found the consolidated complaint plausibly alleges copying and substantial similarity for at least some ChatGPT outputs. That was enough to defeat OpenAI’s motion to dismiss the output-based claims.
The opinion reproduces a detailed summary of A Game of Thrones and an outline for a hypothetical sequel that uses Martin’s characters and setting. The court says a properly instructed jury could find those outputs substantially similar.
“A reasonable jury could find that the allegedly infringing outputs are substantially similar to plaintiffs’ works.” — Sidney H. Stein, U.S. District Judge
“OpenAI’s motion is denied because class plaintiffs’ allegations satisfy the elements of a prima facie claim of infringement as to at least some outputs of ChatGPT.” — Sidney H. Stein, U.S. District Judge
What The Ruling Does Not Decide
The court did not reach a fair use. The opinion states explicitly that nothing in the order should be read as a view on whether the challenged outputs are protected uses of the original works. That question comes later.
Next phases typically involve discovery, expert analysis, and briefing on summary judgment. The timeline will shape when fair-use arguments are decided and whether any claims reach a jury.
How We Got Here And Who Is In The Case
Multiple author lawsuits, including claims by Martin, were centralized this spring in a Manhattan multidistrict litigation, now before Judge Stein. The MDL also includes actions that name Microsoft.
The court’s opinion confirms the MDL posture and addresses the class plaintiffs’ output-based theory. Separate theories tied to model training and data sourcing follow their own paths in the consolidated proceedings.
Why This Matters Beyond One Author
Courts are weighing similar claims across the AI sector. One high-profile case ended in a $1.5 billion settlement with another AI company, underscoring the financial stakes of training and output disputes.
This order raises the odds that a U.S. jury will hear evidence about how model outputs compare to protected books. It also keeps legal pressure on questions of transformation, market harm, and permissions.
Conclusion
The authors cleared a key threshold. The judge found enough in the pleadings and exemplar outputs to let infringement claims proceed. The litigation now turns to facts, experts, and fair-use arguments.
What happens next will shape how courts measure similarity in AI-generated text, and how rights holders and developers negotiate access, licensing, and limits at scale.
📈 Latest AI News
31st October 2025
For the recent AI News, visit our site.
If you liked this article, be sure to follow us on X/Twitter and also LinkedIn for more exclusive content.
