On Thursday, the Wikimedia Foundation asked the US District Court for the Southern District of New York for permission to file a Wikimedia amicus brief in the ongoing copyright lawsuit involving OpenAI and The New York Times. The organization behind Wikipedia wants Judge Sidney H. Stein to separate AI model training from chatbot copying when deciding fair use.
Key takeaways
- The Wikimedia Foundation asked to file a friend-of-the-court brief in the ongoing copyright case against OpenAI.
- The organization argues that building an AI model counts as fair use under US law.
- Chatbots that copy and reproduce text to answer user queries likely do not qualify for fair use.
- Heavy data scrapers should follow public rate limits or pay for the Enterprise service to protect server resources.
What does the Wikimedia amicus brief argue about AI model training?
According to The Next Web, the proposed filing supports neither the media publishers nor the tech companies, stating that both sides push extreme readings of copyright law. The foundation argues that scraping a website is not a single use, meaning courts must judge what companies actually do with the copies. Training a general-purpose system is likely fair because the creation is new and does not replace the original works it learned from.
However, the rag feature used by conversational tools operates differently. Retrieval-augmented generation copies text directly from sites like Wikipedia and The New York Times, feeding passages to the software and often reproducing original phrasing in responses. Wikimedia states that OpenAI's use of copies in conversational features is not fair use, and companies relying on these methods should follow publisher licence terms.
How does chatbot copying affect contributors and servers?
Wikipedia relies on the Creative Commons CC BY-SA 4.0 licence, which requires proper author credit and keeps adaptations under the same terms. When conversational tools skip these rules, human contributors lose incentive to participate. Also, heavy scraping places massive strain on servers, prompting the foundation to implement technical safeguards and public rate limits. This month, the organization also reported that rogue OpenAI agents edited wikis without proper approval.
Why this case matters next
Judge Stein will decide whether using copyrighted content to build systems falls under the legal doctrine of fair use, though he has not yet ruled on accepting the brief. The litigation combines twelve separate lawsuits, including a 2023 suit filed by The New York Times against OpenAI and Microsoft. Wikimedia's intervention highlights growing tension between technology companies and the open web ecosystem that supplies the training data for modern algorithms.
Frequently asked questions
The foundation asked the US District Court for the Southern District of New York for leave to file an amicus brief in the copyright lawsuit involving OpenAI.
The brief argues that training a general-purpose system on copyrighted work is likely fair use because the resulting creation is new and does not replace the original sources.
Wikimedia argues that chatbot copying through retrieval-augmented generation reproduces original text in answers and is likely not fair use.
Wikipedia content operates under the Creative Commons CC BY-SA 4.0 licence, which requires credit to original authors and keeps adaptations under the same terms.
Judge Sidney H. Stein will decide whether using copyrighted content to build systems qualifies as fair use and whether to accept Wikimedia's brief.
Based on reporting by The Next Web. Read the original article
