Wikimedia proposes a use-by-use approach in OpenAI copyright case

AI · First reported 30 Oct, 09:36 · Updated 09:48

Image: Wikimedia Commons · NASA Earth Observatory image by Lauren Dauphin, using Landsat data from the U.S. Geological Survey. Story by Adam Voiland. · Public domain

The Wikimedia Foundation has asked to file a brief in the consolidated copyright litigation involving OpenAI, arguing that courts should assess different AI uses separately. Its proposed position supports fair use for some model training but opposes treating retrieval-augmented generation as fair use when it reproduces source material without respecting licensing terms.

Wikimedia asked the Southern District of New York for permission to file an amicus brief in litigation consolidating 12 lawsuits against OpenAI.
The Foundation argues that qualifying model training may be fair use, while retrieval-augmented generation that draws on source works for readers may not be.
The Foundation says an overly restrictive fair-use ruling could hinder innovation, while blanket fair use could weaken access to secondary sources and encourage publishers to restrict information.

Summary written by Lux AI from 2 articles

Key facts

Filing date: October 29, 2026; Wikimedia announced the request on October 30, 2026.
Consolidated case: 12 lawsuits against OpenAI.
Proposed brief length: 5,831 words, within the court’s 6,000-word limit.
Wikimedia’s requested distinction: Pre-training of ChatGPT’s general-purpose LLM should be fair use; its RAG use should not and should abide by applicable licences.