Wikimedia warns AI copyright ruling could threaten the open web

The Wikimedia Foundation is urging a middle ground in copyright lawsuits involving publishers, OpenAI and Microsoft. In an amicus brief filed Friday, the nonprofit behind Wikipedia argues that courts should distinguish between training large language models (LLMs) on copyrighted works and chatbots reproducing those works for users.

Wikimedia says treating all AI uses alike could either block useful tools or put the open web at risk. It argues that using copyrighted works to pre-train LLMs without a license should qualify as fair use when the models produce something new and different from their sources. But it says AI companies should need a license when systems provide copies in response to user queries in ways that replace engagement with the original works.

The foundation says LLMs can help Wikipedia editors detect vandalism, copyright violations and doxxing, and identify articles that need improvement. It also says services such as ChatGPT reproduce freely licensed Creative Commons text without following the license terms, while AI systems draw visitors away from the sites and communities that maintain the information.

The New York Times alleges that OpenAI infringed its copyright by scraping its website for training data. OpenAI and Microsoft argue that LLM training is fair use. Wikimedia says a ruling should protect access to knowledge while respecting contributors and licensing requirements.