Today (October 27, 2025) in the federal OpenAI ChatGPT Copyright Infringement Litigation case in the Southern District of New York that consolidates fourteen AI copyright cases, Judge Stein in an eighteen-page memorandum refused to dismiss the case at the defendant’s request. In doing so, he made a ruling from which we might read some tea leaves.
Other current federal AI copyright cases such as Thomson Reuters, Kadrey, and Bartz have been dealing with various aspects of copyright law as applied to AI, with some emphasis on the fair use doctrine. One significant discussed aspect is “input-side” infringement, where the focus is on the copying of the plaintiffs’ works, versus “output-side" infringement, where the focus is on the AI output and how it compares with the plaintiffs’ original work.
In today’s ruling Judge Stein held that the plaintiffs subject to the motion had adequately pled their copyright infringement case. As part of that, Judge Stein ruled that the plaintiffs have adequately pled “an output-based infringement claim.” He reasoned that if the facts in the plaintiffs’ pleadings are later found to be true, a jury could find certain allegedly infringing ChatGPT outputs to be substantially similar to the plaintiffs’ original works, using as an example two ChatGPT outputs summarizing or pertaining to plaintiff George R.R. Martin’s novel A Game of Thrones.
The ruling explicitly disclaims that it is dealing in any way with fair use.
Judge Stein’s ruling can be found here:
https://storage.courtlistener.com/recap/gov.uscourts.nysd.641354/gov.uscourts.nysd.641354.617.0.pdf
A list of all the AI court cases and rulings can be found here on Reddit:
[link] [comments]