OpenAI Cites Amicus Briefs to Defend AI Training as Fair Use in NYT Case

OpenAI frames the lawsuit as a threat to fair use

In a post titled "Fair use benefits everyone," OpenAI argues that fair use has long driven innovation, including medical and scientific breakthroughs and new technologies that have helped many people, authors and journalists among them. According to OpenAI, the New York Times is pursuing what it calls an "undeserved payday" and is asking the court to set aside the longstanding principle.

OpenAI says experts and leaders from a range of fields have shared their views with the Court through filed briefs. The post reproduces excerpts "in their own words" and notes that the full filings are available on the public court docket.

What the briefs argue

The excerpts come from a mix of trade groups, library and disability organizations, academics, authors and independent creators. Their arguments differ in focus.

• Purpose and substitution: One excerpt says AI training uses copyrighted works for a fundamentally different purpose and does not give the public substitute access to those works.
• Market dilution: Authors Alliance asks the Court to hold that "market dilution," meaning aggregate competitive pressure from noninfringing outputs, is not cognizable market harm under 17 U.S.C. § 107(4). It says copyright protects authors against substitution for their protected expression, not against the success of new expression that competes with it.
• Wider users of AI: A joint brief from the American Council of the Blind, Association of Research Libraries, American Library Association, Consumer Technology Association and Business Software Alliance says the case affects medical researchers, educators, people with disabilities, farmers, independent creators and sectors from software to law. It says restrictions would affect both institutions building AI and the people using it.
• Startups: Engine and Developers Alliance say the purpose of training a foundation model is not to reproduce the works it learns, but to build a general-purpose tool others can adapt.
• National security: The Foundation for American Innovation argues that the national-security value of capabilities enabled by AI training should be considered under factor four, the assessment of market effects.

Academics, authors and independent creators

Professor Matthew Sag writes that the defendants' use of plaintiffs' works for training is "highly transformative" and strongly favors fair use, because under realistic conditions the models' outputs are in no relevant way similar to the invisible intermediate copies made during training. Joshua S. Gans, a professor of strategic management, argues that copyright supports the incentive to build tools that learn from content as well as the incentive to create it.

Author Kevin Kelly describes each AI model as an entirely new medium for authors. He says a ruling for the plaintiffs would grant far more than the limited monopoly Congress intended, and that otherwise copyright would stifle the creativity it exists to encourage.

Musician and composer Michael Wall and comedian and writer Sarah Rose Siskind, who rely on both copyright protection and AI tools, describe using AI for work around the creative process. Siskind is quoted as saying, "I don't think I could run my business without AI at this point." Wall also doubts his business could survive in its current form without AI.

Public Knowledge points to earlier fair use cases that protected image search, full-text book search (Authors Guild v. Google) and accessible digital books for readers with print disabilities. It says fair use will help ensure AI's benefits reach the public rather than being controlled by a privileged few.

What it means

The post is OpenAI's own presentation of selected excerpts, not a court ruling, and it does not describe the status of the case. It shows how the company and its supporters are framing the dispute: as a question of whether copyright law should extend to AI training, with consequences for libraries, startups, independent creators and users of AI tools. The full filings on the public court docket are the place to read the arguments in context.