The dictionary sues OpenAI | JS

Encyclopedia Britannica and Merriam-Webster have initiated legal proceedings against OpenAI, accusing the AI company of “massive copyright infringement,” as detailed in their complaint.

According to the lawsuit, Britannica, which owns Merriam-Webster, claims ownership of the copyright to close to 100,000 online articles. These articles were allegedly extracted and utilized to train OpenAI’s language models without authorization.

Britannica further accuses OpenAI of breaching copyright laws by producing outputs that include “full or partial verbatim reproductions” of its material and incorporating its articles into ChatGPT’s retrieval augmented generation (RAG) process. This RAG tool allows the language model to search the web or other databases for the latest information when answering queries. Britannica also asserts that OpenAI infringes the Lanham Act, a trademark law, by creating false information and wrongly attributing it to the publisher.

The lawsuit states, “ChatGPT deprives web publishers like [Britannica] of revenue by producing answers to users’ questions that replace, and directly compete with, the content from publishers like [Britannica].” Additionally, Britannica argues that ChatGPT’s inaccuracies threaten “the public’s continued access to high-quality and trustworthy online information.”

Britannica is among several publishers and authors taking legal action against OpenAI over copyright concerns. Other plaintiffs include the New York Times, Ziff Davis (which owns Mashable, CNET, IGN, PC Mag, and others), and more than a dozen newspapers across the US and Canada, including the Chicago Tribune, the Denver Post, the Sun-Sentinel, the Toronto Star, and the Canadian Broadcasting Corporation.

A similar lawsuit by Britannica against Perplexity is still ongoing.

There is currently no strong legal precedent for determining whether the use of copyrighted material to train a language model constitutes copyright infringement. However, in a related case, Anthropic persuaded federal judge William Alsup that using content as training data is sufficiently transformative to be legal. Nonetheless, Alsup ruled that Anthropic broke the law by downloading millions of books illegally instead of purchasing them, resulting in a $1.5 billion class action settlement for the affected authors.

OpenAI did not respond to JS’s request for comment before publication.