Encyclopedia Britannica and its subsidiary Merriam-Webster have initiated legal proceedings against OpenAI in a Manhattan federal court. The lawsuit accuses the AI pioneer of using their proprietary reference materials without authorization to train its large language models.
Britannica vs. OpenAI: A Landmark Copyright Clash
According to the complaint filed on Friday, March 13, 2026, Microsoft-backed OpenAI utilized tens of thousands of online articles, encyclopedia entries, and dictionary definitions to teach ChatGPT how to respond to human prompts. Britannica alleges that this practice has “cannibalized” its web traffic by providing AI-generated summaries that directly substitute for their original content.
An OpenAI spokesperson stated on Monday that their models “empower innovation” and are trained on publicly available data under the framework of fair use. However, Britannica’s legal team argues that the scale of the copying is unprecedented, claiming OpenAI illegally scraped approximately 100,000 articles to feed its GPT models. The lawsuit is seen as a significant addition to the series of high-stakes legal actions brought by authors and news organizations against tech firms.
“Verbatim” Copying and Trademark Concerns
The core of Britannica’s complaint lies in the allegation that ChatGPT often produces “near-verbatim” copies of its encyclopedia and dictionary entries, diverting users who would otherwise visit the publishers’ websites. This diversion directly threatens the advertising and subscription revenue that funds their human-led research and editorial processes.
Furthermore, the lawsuit accuses OpenAI of trademark infringement. Britannica claims that OpenAI creates a false impression of authorization and, perhaps more damagingly, attributes inaccurate AI “hallucinations” to Britannica. This, they argue, deceives users into believing the misinformation is endorsed by the world’s most trusted reference source.
Demands and Ongoing Legal Battles
Britannica is seeking an unspecified amount in monetary damages and a permanent court injunction to block the alleged unauthorized use of its materials. This is not the publisher’s first foray into AI litigation; a similar lawsuit filed by Britannica against Perplexity AI in September 2025 is still ongoing. As courts in the Southern District of New York consolidate over a dozen similar cases, the industry awaits a definitive ruling on the boundaries of “fair use” in the AI era.
What do you think about this legal battle? Should AI models be allowed to use high-quality reference data for free under “fair use,” or should publishers be compensated for the training of these digital assistants? Share your views in the comments!
