OpenAI Highlights Legal Briefs Arguing AI Training Can Qualify as Fair Use

OpenAI on Friday, Oct. 30, highlighted legal briefs arguing that fair use can cover some AI training, adding third-party arguments to an unsettled U.S. copyright debate.

The excerpts came from briefs filed by creators, academics, accessibility advocates and legal experts. Those filers argue that fair use can support some AI training and innovation. OpenAI’s post did not announce a new court ruling.

In June 2025, federal district courts accepted fair-use arguments for model training in Bartz v. Anthropic and Kadrey v. Meta, based on the evidence in each case. The Bartz decision treated pirated books kept in a central library separately from its finding on model training. Neither decision established a nationwide rule.

A May 2025 U.S. Copyright Office report said AI training can be transformative. It also identified unlawful acquisition, model functionality, deployment and effects on markets for original works as relevant factors. TokenPost previously covered OpenAI’s fair-use position in a summary-judgment memorandum and the Justice Department’s separate filing backing that defense.