OpenAI and New York Times clash over fair use in AI training case

Dueling legal briefs set stage for landmark AI copyright ruling

OpenAI, Microsoft, the New York Times, and a coalition of bestselling authors have submitted competing motions for summary judgment in a consolidated copyright lawsuit that could establish the legal boundaries for training large language models on copyrighted content.

The parties filed their briefs Friday with U.S. District Judge Sidney Stein in Manhattan, asking the court to rule on whether AI training constitutes "fair use" under copyright law. According to Reuters, which first reported the filings, the outcome will likely determine the fate of dozens of similar lawsuits against AI companies.

The New York Times sued OpenAI and Microsoft in 2023, alleging the companies used millions of newspaper articles without permission to train ChatGPT. A separate group of authors including John Grisham, Jonathan Franzen, and George R.R. Martin filed similar claims the same year over the use of their books. Those cases were consolidated in New York in 2025.

Competing visions of transformative use

OpenAI argues its training process represents "among the most transformative many of us will see in our lifetimes." The company told the court its pretraining process derives "broad, unprotectable statistical patterns related to language" rather than reproducing protected expression, and maintains its technology has not harmed writers.

Microsoft echoed this position, stating that "neither LLM training nor the use of LLMs in products substitutes for copyrighted books." The company said years of discovery have disproven fears that large language models would displace authors and journalists.

The copyright owners present a starkly different picture. The authors argued that AI is "diluting the market for books across the board" and that fair use doctrine "does not countenance such catastrophic threats to the incentive to create." The news outlets contended that ChatGPT diverts users from their websites and displaces the market for their journalism.

Why it matters

This case will establish whether AI companies can continue training models on copyrighted material without licensing agreements—a practice fundamental to current generative AI development. A ruling against fair use could require AI companies to negotiate licenses with content creators, potentially reshaping the economics of the industry and slowing model development. Conversely, a broad fair use ruling could undermine content creators' ability to control and monetize their work in an AI-driven media landscape.

Divergent precedents from California courts

Judge Stein will navigate conflicting signals from two federal judges in San Francisco who ruled on similar questions in 2025. U.S. District Judge William Alsup called Anthropic's use of books for AI training "quintessentially transformative." Judge Vince Chhabria ruled for Meta two days later, also finding transformative use, but warned that AI training would not qualify as fair use "in many circumstances" and expressed concern that generative AI could "flood the market" with content competing with human creators.

The consolidated case is In re OpenAI Inc Copyright Infringement Litigation, U.S. District Court for the Southern District of New York, No. 1:25-md-03143.

Details of the competing legal briefs were first reported by Blake Brittain at Reuters.