📅 October 30, 2026 • Source: OpenAI
OpenAI published a post titled «Fair use benefits everyone» on October 30, arguing that the court should reject what the company calls an attempt by The New York Times to expand the reach of copyright. In the company's framing, training generative AI models is a transformative, non-expressive and analytical use that does not create substitute access to copyrighted works and does not cause provable market harm. The post is the company's latest public step in a dispute that is being litigated in federal court. The company frames the case as a test of whether copyright can be extended to block the statistical analysis of texts during model training.
The legal setting. The post refers to the consolidated case NYT v. OpenAI and Microsoft before the Southern District of New York. OpenAI emphasizes that authors, journalists, researchers, lawyers, librarians and economists have filed amicus briefs in support of fair use in the case. The company argues that the outcome will reach beyond the parties: it says the ruling will affect medics, educators, people with disabilities and users of open-source AI. The company's 1-billion-plus weekly users figure was claimed in separate OpenAI posts and does not appear in this one. The plaintiffs' position, stated neutrally, is that the use of protected works creates substitute products that compete with the originals.
IMPACT The filings show that the positions submitted to the court are not unified. The Wikimedia Foundation asked the court for leave to file its own amicus brief on October 29 and announced it on October 30: it supports a limited fair use for training while opposing RAG systems that reproduce articles without attribution. For teams building on large models, the case will define whether training on protected texts remains defensible as fair use, and the Wikimedia brief signals that retrieval outputs reproducing articles verbatim may be judged differently from training itself. Publishers, AI vendors and open-source projects all hold a direct stake in how the court frames transformativeness and market harm.
The dispute remains at the briefing stage, and neither the company's post nor the amici filings decides the outcome by itself. For enterprises and developers, the practical near-term effect is continuity: model access and training practices remain unchanged while the case is pending. What the day demonstrates is the shape of the argument: OpenAI presents broad professional support for fair use, while the Wikimedia filing shows that even supporters of training exemptions draw a line at output-side reproduction. Further briefs and the court's treatment of the substitute-market question will indicate where the case is heading.
Sources: openai.com and engadget.com
Continue the conversation: https://t.me/AIAutomationArchitecture · Bluesky
Top comments (0)