The Wikimedia Foundation is asking a federal court to distinguish between using copyrighted material to train AI and reproducing that material in chatbot answers.
The nonprofit behind Wikipedia announced the filing on October 30, 2026. Its proposed amicus brief, a submission offering an outside organization’s perspective, supports neither side.
The New York Times sued OpenAI and Microsoft in December 2023, alleging that they used its journalism without permission.
OpenAI disputes the claims and argues that training AI on publicly available material qualifies as fair use. Fair use allows certain uses of copyrighted work without the owner’s permission.
The case could affect publishers’ ability to control how AI companies use their work and whether those companies need a license.
Why Wikimedia is weighing in on the OpenAI copyright case
Wikimedia argues that an overly restrictive ruling could make developing AI tools too expensive for open-source developers. But giving AI companies blanket protection could encourage publishers to restrict access to their content, leaving fewer sources available to Wikipedia editors.
Wikipedia volunteers rely on journalism and other published research to write and check articles. AI developers then use Wikipedia as a source for their models.
Wikimedia also uses machine-learning tools to help editors detect vandalism and copyright violations. The Foundation wants to keep those tools available without losing access to the sources Wikipedia needs.
Training a general-purpose model can qualify as fair use, it argues, when the result is something new and different from the original works. Reproducing those works in chatbot answers raises a different issue, particularly when the answer replaces a visit to the source.
The brief notes that some services use material they own or have licensed. Whether a particular use is fair use depends on the circumstances.
For Wikipedia, the Foundation wants AI companies retrieving and reproducing its content to follow its licensing requirements. Those include crediting contributors and sharing adaptations under the same Creative Commons license.