Home
Latest reporting
News
OpenAI published excerpts from briefs by creators, academics and legal experts supporting fair use for AI training in U.S. copyright litigation.
OpenAI on Oct. 30, 2026 said creators, academics, accessibility advocates, legal experts, librarians, economists and business leaders have filed briefs backing fair use for AI training in U.S. copyright litigation, framing the issue as a test of how copyright law applies to generative AI. In its post, the company said the New York Times lawsuit asks the court to set aside what it described as a longstanding fair-use principle, while supporters argue that training large language models is legally distinct from copying works for substitution.
Evidence and context
The company published excerpts from briefs filed by groups and individuals including the American Council of the Blind, Association of Research Libraries, American Library Association, Consumer Technology Association, Business Software Alliance, Authors Alliance, Engine and Developers Alliance, Foundation for American Innovation, Public Knowledge, Kevin Kelly, Matthew Sag, Ron Suskind, Brian L. Frye and the United States government. The excerpts, attributed by OpenAI to those filers, argue that AI training serves a different purpose from the underlying works, does not provide substitute access to copyrighted material, and does not create the kind of market harm copyright law is designed to prevent.
Several filings, as quoted by OpenAI, emphasized that fair use has historically supported tools such as image search, full-text book search, accessible digital books and online publication of legal standards. Others argued that AI tools can aid independent creators and businesses by reducing operational costs and expanding creative output. OpenAI also quoted the U.S. government as saying that narrowing fair-use doctrine to exclude LLM training would be inconsistent with copyright principles and could hinder scientific and economic progress.
The post comes after a series of 2025 rulings and settlements that have shaped the legal backdrop for AI training disputes. On June 25, 2025, a court in Kadrey v. Meta found AI training fair use on the evidentiary record presented, according to a Sullivan & Cromwell analysis published June 26, 2025. On June 23, 2025, a Crowell analysis published June 30, 2025 said Bartz v. Anthropic distinguished fair-use AI training from the acquisition of pirated books, and a court memorandum filed Oct. 17, 2025 preliminarily approved a $1.5 billion settlement over pirated-book claims in that case, according to the enrichment provided here.
Those rulings do not establish a universal exemption for AI training, and the legal questions remain fact-specific under U.S. fair-use law. The source material also reflects a broader split in the litigation landscape, with some courts and analyses treating training, source-copy acquisition, retention and output distribution as separate copyright issues rather than a single category. OpenAI did not provide a next dated court event in the supplied material.
