Wikimedia Foundation urges balanced fair use in NYT’s copyright lawsuit against OpenAI

Wikimedia Foundation’s stance on fair use

The Wikimedia Foundation has filed an amicus brief in the copyright lawsuit brought by The New York Times against OpenAI and Microsoft. The foundation argues for a nuanced application of fair use principles in AI model training and deployment, warning that overly restrictive or overly broad interpretations could have unintended consequences for innovation and the open web.

According to the foundation, fair use should apply to specific scenarios, such as training AI models on copyrighted content, but only if the output is transformative and distinct from the original sources. It also supports the use of retrieval-augmented generation (RAG) under fair use, provided there is proper attribution.

Potential impact of the lawsuit

The New York Times’ lawsuit centers on whether OpenAI and Microsoft’s use of published content for AI training violates copyright law. The newspaper argues that allowing companies to freely use such content could undermine journalism and creative industries. OpenAI and Microsoft counter that their practices are essential for technological advancement and fall under fair use.

Wikimedia’s brief highlights the risk of a blanket fair use ruling, which could lead to a "closed web" where access to secondary source materials is restricted, weakening the foundation of collaborative platforms like Wikipedia. Stephen LaPorte, General Counsel for the Wikimedia Foundation, stated that Wikipedia thrives on open and accessible human knowledge, and a ruling that limits this could stifle its collaborative process.

A summary judgment in the case is expected soon, and the outcome could set a legal precedent for how AI companies use publicly available data.

Earlier Wikimedia intervention in the case

This is not the first time the Wikimedia Foundation has weighed in on the lawsuit. In an earlier amicus brief, the foundation distinguished between two uses of copyrighted content: training AI models and generating responses using RAG. It argued that training on such content is likely fair use, but using it in RAG responses without attribution may not be. The current brief expands on this position, emphasizing the need for clarity and balance in fair use interpretations.

What this means

Lazyfounder analysis — our interpretation, not reported fact.

The Wikimedia Foundation’s intervention in this case underscores the high stakes for AI development and the open web. A ruling favoring The New York Times could force AI companies to renegotiate licensing agreements with publishers, potentially increasing costs and limiting access to training data. Conversely, a broad fair use ruling could embolden tech companies to use copyrighted content without constraints, raising concerns about fairness and sustainability for content creators.

For founders and operators in the AI space, this case is a bellwether. A restrictive ruling could require rethinking business models, particularly for startups that rely on scraping or licensing data. It may also accelerate the shift toward synthetic data or partnerships with content providers. The outcome could also influence how platforms like Wikipedia are used in AI training, shaping the future of open knowledge ecosystems.

Related on Lazyfounder

• Wikimedia Foundation
• The New York Times
• OpenAI
• Microsoft
• More technology news
• Wikimedia Foundation argues AI training is fair use but RAG responses are not in OpenAI lawsuit