Wikimedia Foundation Urges Nuanced Fair Use Ruling in NYT v. OpenAI Case

The Wikimedia Foundation has entered the copyright lawsuit The New York Times filed against OpenAI and Microsoft, submitting an amicus brief that calls for a careful, middle-ground approach.

The Times argues that companies like OpenAI should not have free rein over its published content, while OpenAI and Microsoft contend that training and deploying AI models on published data is necessary for technological development and falls under fair use.

The Wikimedia Foundation's brief sides partly with the tech companies, warning that an overly restrictive reading of fair use could inhibit innovation. But it also cautions that blanket fair use could produce a closed web where knowledge suffers from a "lack of available secondary source materials."

Instead, the foundation proposes applying fair use to two specific cases: certain model training scenarios and retrieval-augmented generation (RAG), a technique that improves large language model output by referencing authoritative knowledge bases outside its training data. The foundation does not support RAG when a model regurgitates existing articles without attribution. On training, it argues that scraping copyrighted content should count as fair use when the resulting models are "something new and different from the original sources."

"A ruling that does not apply fair use principles carefully could threaten the sustainability of free knowledge, and stifle the development of open source technologies, ultimately impacting access to information for everyone in both cases," General Counsel Stephen LaPorte wrote, adding that the court needs to understand Wikipedia is built through a collaborative process requiring "a strong foundation of open and accessible human knowledge."

A summary judgment is expected soon. The case is one of more than a dozen similar lawsuits working through the courts, but it could set an important precedent.

Source: Engadget