Wikimedia warns AI copyright ruling could jeopardize open web and Wikipedia

In an amicus brief filed Friday, the Wikimedia Foundation joined the legal fight surrounding The New York Times’ lawsuit against OpenAI and Microsoft, warning that an overly simplistic copyright ruling could endanger the open web The Verge. The nonprofit argues that training large language models (LLMs) on copyrighted material should be treated as fair use, while chatbots that reproduce that material to users should not.

Wikimedia says Wikipedia’s data is “one of the most critical sources of data for all artificial intelligence and machine learning training,” yet the same systems siphon away visitors, weakening the volunteer communities that maintain the sites. The filing notes that services like ChatGPT reproduce Creative Commons-licensed text without honoring the license terms.

The foundation proposes a compromise: scraping for training should be permissible, but AI outputs that directly replace original works must respect licensing and attribution requirements. CEO Bernadette Meehan stresses that expanding AI access must protect contributors and the collaborative model that underpins free knowledge.

Our read: Wikimedia’s stance highlights a rare middle ground—protecting the free-knowledge ecosystem while allowing transformative AI training—yet the outcome will hinge on how courts balance fair-use doctrine against publisher demands.