Wikimedia Foundation Tries to Save Itself From Oblivion by Brokering AI Copyright Deal

Why it matters

The Wikimedia Foundation has entered the copyright fight between the New York Times and OpenAI and Microsoft, filing an amicus brief that rejects both sides' positions as overly sweeping. The Times alleges the companies illegally used its copyrighted articles to train chatbots that now rival it as an information source, while the defendants claim web scraping for training ChatGPT and Copilot is fair use. Wikimedia, which runs Wikipedia and other collaboratively edited sites, proposes a middle path: training on freely available online content should be permitted, but reproducing expressive material in chatbot outputs should not. The brief acknowledges that large language models can serve the nonprofit's mission, such as helping editors and detecting personal information, but warns that a careless fair-use ruling could undermine the collaborative, license-based publishing model Wikipedia depends on. The stakes extend beyond this case, since courts are still untangling how opaque AI systems handle training data, and a final resolution could take years across the many publisher lawsuits now pending.