Overview
- The publishers filed their complaint in the Southern District of New York on March 13, alleging OpenAI copied nearly 100,000 Britannica articles and Merriam-Webster entries to train GPT models.
- They say ChatGPT reproduces curated reference content near verbatim, citing examples like an identical definition of “plagiarize” and a Britannica article’s exact selection and ordering of quotes about the Hamilton–Burr duel.
- The lawsuit alleges a second use of their works through ChatGPT’s retrieval‑augmented generation workflow, which can pull Britannica material in real time when answering queries.
- Trademark claims under the Lanham Act accuse OpenAI of attaching the publishers’ names to hallucinated or incomplete text in ways that could mislead users and harm their reputation.
- Britannica and Merriam‑Webster seek damages, disgorgement of profits, and a permanent injunction, while OpenAI disputes wrongdoing, saying its models use publicly available data and are grounded in fair use; a prior Britannica suit against Perplexity remains pending.