Britannica and Merriam-Webster Sue OpenAI for Copyright Infringement Over Unauthorized Use of Nearly 100,000 Articles for AI Training

Encyclopedia Britannica and its subsidiary Merriam-Webster have filed a lawsuit against OpenAI in Manhattan federal court, alleging massive copyright infringement. The complaint claims that OpenAI unlawfully used nearly 100,000 of Britannica's online articles to train its large language models, including ChatGPT.

The lawsuit argues that OpenAI's actions constitute direct infringement of Britannica's copyrights, as the AI models allegedly reproduce or closely paraphrase Britannica's content without authorization. Furthermore, the suit contends that OpenAI's use of retrieval-augmented generation (RAG) systems, which pull information from Britannica's content in real-time to generate responses, exacerbates the infringement.

A key argument is that ChatGPT's ability to provide comprehensive answers to user queries serves as a substitute for visiting Britannica's websites, thereby depriving them of subscription and advertising revenue essential for funding the creation of high-quality, reliable content.

The lawsuit also includes claims of trademark infringement under the Lanham Act, alleging that ChatGPT sometimes generates fabricated information and falsely attributes it to Britannica, potentially damaging the publisher's reputation for accuracy and trustworthiness. The plaintiffs are seeking monetary damages and an injunction to prevent OpenAI from continuing to use their copyrighted materials.

This lawsuit is part of a growing trend of legal challenges against AI companies by content creators and publishers, including The New York Times and Ziff Davis, who are concerned about the unauthorized use of their intellectual property for training AI models. OpenAI has disputed the claims, stating that its models are trained on publicly available data and that their use is protected under the fair use doctrine.