Source: Diff

The Wikimedia Foundation has submitted an amicus brief in New York federal court in In re: OpenAI, Inc. Copyright Infringement Litigation, concerning AI uses of copyrighted material. Stephen LaPorte, the Foundation’s General Counsel, explains the submission and presents the Foundation’s arguments about protecting access to knowledge. The litigation combines 12 lawsuits, including The New York Times Co. v Microsoft and OpenAI. The New York Times alleges that OpenAI and Microsoft used the newspaper’s copyrighted content without permission; publishers challenge acquisition, training and subsequent uses in AI services. OpenAI and Microsoft argue that using publicly available material to develop AI systems qualifies as fair use, while OpenAI advances a broader defense covering development and operation. The Foundation says its submission supports neither OpenAI nor The New York Times and instead asks the court to consider effects on open knowledge. LaPorte explains that Wikipedia volunteers depend on journalism and research, while AI systems also draw on Wikipedia’s content.

The Foundation proposes examining different stages of AI use separately rather than applying one conclusion to every use. The Foundation argues that training may qualify as fair use when the resulting model differs meaningfully from the copyrighted works used. For retrieval-augmented generation, which uses retrieved information to construct answers, the Foundation argues that retrieving and reproducing copyrighted content should respect source licensing requirements. For Wikipedia content, the Foundation calls for attribution and compliance with applicable share-alike requirements. LaPorte warns that excluding training from fair use could put AI development beyond the resources of many open-source developers and hinder tools used by Wikimedia volunteers. Conversely, LaPorte argues that blanket protection for AI uses could encourage publishers to restrict access and weaken the attribution and visibility supporting knowledge-producing communities. The Foundation also calls on companies collecting or retrieving Wikipedia content to respect public API rate limits or use Wikimedia Enterprise API offerings.