Britannica sues OpenAI for scraping nearly 100,000 articles

Encyclopedia Britannica — which also owns Merriam-Webster — is suing OpenAI for scraping nearly 100,000 copyrighted articles to train its AI models, TechCrunch reports. The lawsuit makes three claims: OpenAI trained on the content without permission, ChatGPT generates full or partial verbatim reproductions of the articles, and OpenAI uses the articles in ChatGPT's retrieval-augmented generation workflow.

There's also a trademark angle. Britannica alleges Lanham Act violations when ChatGPT produces hallucinated information and falsely attributes it to Britannica, presenting made-up facts as if they came from the encyclopedia.

"ChatGPT starves web publishers like [Britannica] of revenue by generating responses that substitute, and directly compete with, the content from publishers," the complaint states. Britannica joins the New York Times, Ziff Davis, Chicago Tribune, Denver Post, Toronto Star, and CBC in the growing line of publishers suing OpenAI.

In a related case, Anthropic convinced Judge Alsup that training on copyrighted material constitutes transformative, legal use — but the court also ruled Anthropic illegally downloaded books, leading to a $1.5 billion class action settlement.

Previously:

• New York Times sues OpenAI and Microsoft, claiming copyright infringement
• Copyright holders are finally beginning to push back against OpenAI