OpenAI promotes fair-use arguments in its New York Times dispute
OpenAI published excerpts from supporting fair-use briefs on October 30 in its dispute involving The New York Times. The arguments concern whether training on copyrighted material requires permission, a question that affects developers' data sourcing and publishers' licensing businesses.
Spec sheet
- Update
- 30 October 2026
- Parties
- OpenAI and The New York Times
- Material
- Excerpts from supporting court briefs
- Neutral background
- District court opinion dated 4 April 2025
OpenAI's selection includes Authors Alliance, Public Knowledge and the United States government. The company argues that fair use supports AI training. The court's April 2025 opinion records the publishers' opposing account: their works were used in training and reproduced in outputs, potentially substituting for visits to their websites. That puts subscription, advertising and licensing revenue at the center of the disagreement.
Training and generated answers present different questions for a software team. A decision about one use would need to be read for its scope before being applied to another. Developers should keep separate records of training inputs, retrieved material and displayed output, so a change in one area does not leave the whole product's data history tangled together.
Claim check
Sources
- Fair use benefits everyone, October 30 · OpenAI · 30 Oct
- The New York Times lawsuit · OpenAI · 31 Oct
- Opinion, April 4, 2025, The New York Times Co. v. Microsoft et al., 23-cv-11195 (SHS), pp. 5, 7–9 and 20–24 · U.S. District Court, Southern District of New York · 4 Apr