Two questions, two answers
The Wikimedia Foundation has asked the judge in The New York Times' copyright lawsuit against OpenAI to split the issue in two. The foundation, which runs Wikipedia, filed an amicus brief with the court on Thursday.
Its core argument: training an AI model on copyrighted content is likely covered by fair use. A chatbot copying content from websites to answer a user's question, however, likely is not.
Why it matters
The distinction separates two different practices at the heart of the AI copyright debate: the training process behind a model and the reproduction of content when generating an answer for a user. If the court adopts this split, it could set a benchmark for this case and similar copyright disputes.
Wikimedia's position
The foundation does not oppose the use of open knowledge sources in AI training. But it argues that systems which reproduce source sites' content verbatim while answering user questions can undermine copyright holders' rights. Details of the brief and the case's next steps are not yet clear.
