Encyclopedia Britannica and Merriam Webster have filed a lawsuit against OpenAI, accusing the company of large-scale copyright infringement tied to how it trains and operates its artificial intelligence systems.
The publishers argue that OpenAI used copyrighted material from their online encyclopedia and dictionary entries without permission while building its language models, and they say the AI company now produces responses that compete directly with their original content.
According to the complaint filed by Encyclopedia Britannica, which owns Merriam Webster, OpenAI allegedly scraped nearly 100,000 copyrighted online articles and used them to train its large language models.
The publisher says the AI company reproduced “full or partial verbatim” passages from its work and integrated that material into ChatGPT’s retrieval augmented generation workflow, which scans online databases to provide updated answers.
Publishers say AI harms trusted information
Britannica argues that ChatGPT reduces traffic to trusted publishers by answering user questions directly with AI generated responses. The lawsuit claims the system “starves web publishers of revenue” because users receive answers from the chatbot instead of visiting the original sources that created the information.
The company also says some AI responses contain hallucinated content that incorrectly credits Britannica, which the lawsuit claims violates trademark law under the Lanham Act. The publisher warns that such errors threaten “the public’s continued access to high quality and trustworthy online information.”
Britannica now joins several media companies and publishers that have already taken legal action against OpenAI over copyright use in AI training. The New York Times, Ziff Davis, and several major North American newspapers have filed similar lawsuits.
Britannica has also filed a separate case against Perplexity, which remains pending in court.
