OpenAI Highlights Briefs Supporting Fair Use in AI Training

OpenAI highlighted support from legal experts, creators, and organizations for applying the legal principle of fair use to AI model training. On October 30, 2026, the company published excerpts from amicus briefs filed in public court dockets related to ongoing litigation with The New York Times.

A joint filing by the American Council of the Blind, Association of Research Libraries, American Library Association, Consumer Technology Association, and Business Software Alliance stated that fair use affects institutions building AI as well as communities using the technology, including medical researchers, educators, and independent creators. The group noted that many depend on open source AI systems developed by others rather than training models themselves.

The Authors Alliance asked the court to hold that market dilution, or aggregate competitive pressure from noninfringing outputs, is not cognizable market harm under 17 U.S.C. § 107(4). Professor of Law Matthew Sag submitted that training uses are transformative because outputs under realistic conditions are not similar to intermediate copies made during training. Professor of Strategic Management Joshua S. Gans stated that treating every displaced sale as actionable could weaken incentives to build learning tools without reliably protecting content creation incentives.

In its filing, the United States Government stated that narrowing fair use to exclude large language model training would conflict with basic copyright law principles under U.S. CONST. art. I, § 8, cl. 8. Additional submissions came from Engine and Developers Alliance, Foundation for American Innovation, Public Knowledge, Ron Suskind, Brian L. Frye, Kevin Kelly, Michael Wall, and Sarah Rose Siskind.

Public Knowledge cited past fair use rulings, including decisions from 2007, 2014, 2015, and 2023, to illustrate how fair use has protected tools that navigate large information collections and provide accessible digital books.