The Wikimedia Foundation has filed an amicus brief in the consolidated copyright litigation against OpenAI in federal court in New York, asking Judge Sidney H. Stein to treat training a general-purpose AI model differently from chatbots that copy and reproduce source text. The nonprofit behind Wikipedia sided with neither The New York Times nor OpenAI and Microsoft, arguing that both sides advance "overbroad and simplified" readings of fair use.
The case, In re: OpenAI, Inc. Copyright Infringement Litigation, sits in the US District Court for the Southern District of New York and combines 12 lawsuits against OpenAI, including the Times' 2023 suit against OpenAI and Microsoft. Wikimedia requested leave to file the brief on Thursday and announced it on Friday. Stein has not ruled on fair use or on whether to accept the brief, and a summary judgment in the case is expected soon.
“We are taking a position in this case to defend the free and open human knowledge that Wikipedia depends on.”— Bernadette Meehan, Chief executive of the Wikimedia Foundation
On training, the Foundation lines up with the AI companies' broader argument. The brief says training a general-purpose model is "likely to be fair" because the resulting model is new and does not replace the works it learned from. Wikimedia also warns that an overly restrictive reading of fair use could hold back technological innovation, including tools that benefit open projects like Wikipedia.
Key facts
- 01Wikimedia filed an amicus brief in In re: OpenAI, Inc. Copyright Infringement Litigation, which consolidates 12 lawsuits.
- 02The brief says training a general-purpose AI model is "likely to be fair" use under copyright law.
- 03Wikimedia argues RAG systems that reproduce text should follow publisher licenses such as Wikipedia's CC BY-SA 4.0.
- 04Judge Sidney H. Stein has not yet ruled on fair use or on whether to accept the brief.
- 05Wikipedia hosts more than 67 million articles in 300 languages.
The brief asks the court to judge each use separately rather than issue a single verdict on AI, citing the Supreme Court's 2023 Warhol ruling. It also argues that scraping a website is not, by itself, a "use" under copyright law.
Retrieval-augmented generation is where Wikimedia draws its line. RAG systems pull text from sites such as Wikipedia or the Times to answer a user's question, and the brief says those answers often reproduce much of the original. Wikimedia argues that companies using RAG and similar "grounding" techniques should follow each publisher's license terms.
For Wikipedia, those terms are the Creative Commons CC BY-SA 4.0 license, which requires credit and that adaptations be shared under the same license. The Foundation also warns that a blanket fair-use ruling could push the web toward closed publishing, leaving less secondary source material available. Wikipedia hosts more than 67 million articles in 300 languages.
“A ruling that does not apply fair use principles carefully could threaten the sustainability of free knowledge”— Stephen LaPorte, General counsel of the Wikimedia Foundation
Server load is part of the argument. The brief says heavy scraping strains Wikimedia's infrastructure, and the Foundation wants high-volume users to follow its public rate limits or pay for its Wikimedia Enterprise service. Earlier this month, Wikimedia said rogue OpenAI agents had edited its wikis without approval.
Stephen LaPorte, the Foundation's general counsel, framed the filing around nonprofit knowledge projects, saying any conversation about fair use and copyright in the age of AI "cannot be limited to for-profit organizations." He added that the court needs to understand Wikipedia is created through a collaborative process.
The parties remain far apart. OpenAI and Microsoft argue that training on publicly available data is fair use, while the Times alleges that using its content for any purpose, including training, infringes its copyrights. Neither OpenAI, Microsoft nor the Times has publicly responded to the Wikimedia brief.
The consolidated New York case is one of more than a dozen similar suits in US courts. Nearly 400 local newspapers sued OpenAI and Microsoft in June, and in September the Justice Department told a New York court that AI training is fair use.
Working on something we should cover, or seeing a story we missed? Send leads, documents, or feedback to hello@aichatdaily.com. For sensitive tips, see our secure tips page for Signal and PGP options.
Spotted an error? Email hello@aichatdaily.com with the URL and the issue, or read our full corrections policy.
