Encyclopædia Britannica Sues OpenAI: ChatGPT Accused of Plagiarism

Britannica and Merriam-Webster at War with OpenAI

Encyclopædia Britannica and dictionary publisher Merriam-Webster have filed a lawsuit against OpenAI. They accuse the company of using their copyrighted content to train its artificial intelligence model, ChatGPT, and generating responses that are "substantially similar" to their original content.

According to the plaintiffs, OpenAI has repeatedly copied Britannica's content without obtaining prior permission. They claim that GPT-4, one of OpenAI's models, has "memorized" a significant amount of protected content and is capable of producing nearly identical copies of substantial portions of their texts on demand.

Troubling Examples of Similarities

The complaint filed by Britannica includes concrete examples where the responses generated by OpenAI's models are compared with the original text from Britannica. In these examples, entire passages appear to match word for word, raising questions about the unauthorized use of this content.

Britannica also argues that this practice has "cannibalized" its web traffic. Indeed, instead of directing users to Britannica's site, the responses generated by ChatGPT directly substitute or compete with the encyclopedia's content, which is not the case with a traditional search engine.

A Context of Growing Litigation

This case is part of a broader trend of copyright lawsuits against artificial intelligence companies. The New York Times has also filed a similar complaint against OpenAI, accusing the company of copying significant amounts of its protected content. In September, Anthropic settled a class action for using copyrighted books to train its models, resulting in a payment of $1.5 billion to the affected authors.