The New York Times, OpenAI, and Microsoft presented motions for summary judgment to a federal judge on Friday, advancing the 2023 copyright lawsuit into a phase that could define whether scraping published works to train large language models constitutes fair use2.
District Judge Sidney Stein is expected to rule in the coming months on whether the case, or key portions of it, should move to trial next year.
The Times argued that OpenAI and Microsoft copied its works at scale to build commercial substitutes, which it believes violates the law. The newspaper contended the case should be evaluated against binding copyright precedents established by the Supreme Court and the Second Circuit rather than recent lower court rulings.
OpenAI countered that existing copyright laws and legal precedents allow tech companies to scrape publicly available works to train their large language models, and that reported facts are not protected by copyright law. OpenAI cited studies suggesting that ChatGPT outputs do not create commercial substitutes of the Times' work. Microsoft made similar arguments for its Copilot AI product.
OpenAI's filing cited two recent California federal court decisions, Kadrey v. Meta and Bartz v. Anthropic, which found that using copyrighted works to train AI models constituted fair use because the training was considered transformative. The Times responded that those cases are different and cannot be applied to this suit because those plaintiffs did not show that AI outputs competed with their products.
The Justice Department filed a brief on Thursday supporting OpenAI, arguing that training large language models from millions of digital works delivers a transformative public benefit that offsets any competitive harm to the publishing industry.
ANALYSIS The DOJ brief adds federal executive-branch weight to OpenAI's fair-use defense at a moment when the court is weighing whether to resolve the case before trial. The divergence between the Times' framing, centered on commercial substitution, and OpenAI's framing, centered on transformative purpose, sets up the core legal question the court must resolve.
Axios, which reported the filing details, has a licensing and technology agreement with OpenAI that allows OpenAI to access part of Axios' story archives and helps fund Axios' expansion into several local cities.