Britannica sues OpenAI
Encyclopedia Britannica filed a lawsuit against OpenAI over AI training data, adding to copyright and legal headwinds for model training workflows. Enterprises are already factoring legal risk and auditability into data and model‑training choices. (thehindu.com)
Complaint was filed on March 13, 2026 in the U.S. District Court for the Southern District of New York and assigned case number 1:26‑cv‑02097. (documentcloud.org) Plaintiffs are Encyclopædia Britannica, Inc. and Merriam‑Webster, Inc., and the defendants named include OpenAI, Inc.; OpenAI LP; OpenAI GP LLC; OpenAI LLC; OpenAI OpCo LLC; OpenAI Global LLC; OAI Corporation, LLC; OpenAI Holdings, LLC; and OpenAI Group PBC, with the complaint demanding a jury trial. (documentcloud.org) The complaint alleges OpenAI copied nearly 100,000 of Britannica’s online articles to train its GPT models and that ChatGPT generates responses that contain “verbatim or near‑verbatim reproductions, summaries, or abridgements” of plaintiffs’ works. (thenextweb.com) Britannica brings claims under the Copyright Act and the Lanham Act, and asks the court for unspecified monetary damages plus injunctive relief to stop the alleged copying and misleading attributions. (usnews.com) The filing explicitly flags use of plaintiffs’ content in retrieval‑augmented generation (RAG) workflows and echoes a separate lawsuit Britannica filed against Perplexity in 2025, saying the full scope of scraping and reuse is known only to OpenAI. (techstartups.com) OpenAI’s public response reiterated that its models are trained on “publicly available data” and that those uses are “grounded in fair use,” according to statements reported after the complaint was filed. (usnews.com) The complaint notes that plaintiffs retain copyright in their nearly 100,000 online articles but acknowledges that precise quantities and indexing practices are within OpenAI’s control, a point the suit uses to justify discovery into training datasets and related systems. (thenextweb.com)
