Britannica and Merriam-Webster Sue OpenAI for Massive Plagiarism

Britannica and Merriam-Webster at War with OpenAI

Encyclopedia Britannica and Merriam-Webster have decided to sue OpenAI, filing a complaint in federal court in New York. This action follows a similar lawsuit filed against Perplexity six months ago. The two publishers accuse ChatGPT of using their content without permission, first for training its models and then by reproducing this content in its responses.

Britannica claims copyright over approximately 100,000 online articles as well as the printed volumes of the New Encyclopedia Britannica. The complaint is based on two legal grounds: violation of the Copyright Act of 1976 and trademark infringement under the Lanham Act.

Three Levels of Accusations

The complaint details three levels of infractions. First, Britannica's articles are alleged to have been integrated into the early versions of ChatGPT through databases like "WebText" for GPT-2 and "Common Crawl" for GPT-3 and its successors. These corpora explicitly include the sites of Britannica and Merriam-Webster.

Second, the RAG (Retrieval Augmented Generation) system is implicated. This process allows ChatGPT to consult sources in real-time to complete its answers, including those from the plaintiffs, without a licensing agreement.

Finally, the complaint includes screenshots showing that GPT-4 reproduces passages from Britannica articles verbatim, notably "Education" and "Tourism." Ironically, ChatGPT even quoted the definition of the verb "plagiarize" from the Merriam-Webster Collegiate Dictionary. In some cases, the model omits paragraphs without indicating this, thus distorting the original content.

Economic and Reputational Consequences

Since discontinuing its print edition in 2012, Britannica has focused on digital, deriving its revenue from subscriptions and advertising, directly linked to traffic on its site. When ChatGPT answers a question using a Britannica article without redirecting to the site, it deprives the encyclopedia of crucial visits, capturing the value of the content without compensation.

The complaint also addresses the issue of ChatGPT's "hallucinations." When the AI invents information and attributes it to "Britannica" or "Merriam-Webster," it violates the Lanham Act. These false attributions can mislead users about the origin of the content and harm the reputation of brands established for over 250 years.

In September 2025, Britannica and Merriam-Webster had already sued Perplexity for similar reasons, a case that is still ongoing. To date, more than 90 copyright infringement lawsuits have been filed against AI companies in the United States. The most notable case remains that of Anthropic, which settled a lawsuit for $1.5 billion in 2025 after being accused of using pirated books to train its models.