OpenAI: seven professors challenge blanket fair use for books

According to ETLegalWorld’s article “US Law Professors Challenge OpenAI's Fair-Use Defense for AI Training”, seven US law professors have asked the federal court for the Southern District of New York to reject a blanket finding of fair use for training AI models on copyrighted books. The publication reports that their amicus curiae brief was filed on November 6, 2026, backing the plaintiffs’ summary judgment request and opposing the defendants’ request in the OpenAI litigation. The brief has not been examined directly; the account of its contents and procedural status relies exclusively on the news report.

In ETLegalWorld’s account, the professors read Section 107 of the US Copyright Act as requiring scrutiny of the purposes and consequences of individual uses. Their approach separates how books are obtained and copies stored or shared from their role in building an AI system and the system’s later deployment. They also draw a distinction between tools serving translation or accessibility needs and commercial systems producing works that could compete with the books.

On economic consequences, ETLegalWorld attributes to the signatories concerns that copying without permission could weaken the training licensing market and that generated content could affect sales of original books. The professors consequently seek scrutiny of the particular conduct challenged, rather than a conclusion covering AI training as a whole. They also ask the court to decide in the plaintiffs’ favour on specified instances of allegedly unauthorised downloading, reproduction and circulation, where uncontested facts permit that outcome. These remain the signatories’ requests and projected risks: at the time of the article, the court had not decided the arguments in the brief.