Wikimedia Foundation urges courts to split AI copyright rulings on training versus outputs

The Wikimedia Foundation, the nonprofit behind Wikipedia, has filed an amicus brief in a group of landmark AI copyright cases. It argues that large language models should be legally protected when trained on copyrighted material, but not when chatbots reproduce that material for users. The foundation warns that an overly black-and-white ruling could either cut off useful AI tools for Wikipedia editors or threaten the open web. Its position is neutral, taking neither side in lawsuits brought by publishers including The New York Times against OpenAI and Microsoft.