Copyright: Encyclopaedia Britannica sues OpenAI over AI training
The famous online encyclopedia accuses the ChatGPT developer of copyright and trademark infringements and is suing for damages and injunctive relief.
(Image: Shutterstock/ioda)
The online encyclopedia Encyclopaedia Britannica and its subsidiary Merriam-Webster, the oldest dictionary publisher in the USA, have sued ChatGPT developer OpenAI in a US District Court in Manhattan. They accuse the US corporation of “massive copyright infringement” in the training of its AI models. This was reported by the Reuters news agency.
OpenAI allegedly used nearly 100,000 online articles as well as encyclopedia and dictionary entries from Encyclopædia Britannica and Merriam-Webster without authorization to train its chatbot ChatGPT, according to the lawsuit filed on Friday (Case No. 1:26-cv-2097). According to the lawsuit, ChatGPT has created near word-for-word copies of Britannica's encyclopedia entries, dictionary definitions, and other content, thereby diverting users who would otherwise visit Britannica's websites. “ChatGPT deprives website operators like the plaintiffs [Britannica, ed.] of revenue by generating answers to user queries that replace the content of these operators and are in direct competition with them. To create these substitute products, the defendants mass-copy copyrighted content of the plaintiffs and other website operators without their permission or compensation,” the lawsuit states.
In addition to copyright infringement, Britannica also accuses OpenAI of violating a trademark law (Lanham Act) “by generating fabricated content or 'hallucinations' and falsely attributing them to the plaintiffs.” It further states: “Similarly, the defendants violate the plaintiffs' trademark rights under the Lanham Act by ChatGPT misleadingly omitting parts of the plaintiffs' content without disclosing these omissions, and displaying the incomplete and erroneous reproductions alongside the plaintiffs' well-known trademarks.” ChatGPT endangers “the public's continued access to high-quality and trustworthy online information,” the accusation reads.
In the lawsuit, Britannica seeks compensation in an unspecified amount as well as an injunction to prevent the alleged trademark infringement. According to Reuters, Britannica filed a similar lawsuit against the AI start-up Perplexity AI last year, which is still pending.
Videos by heise
Britannica's action against OpenAI is one of numerous legal disputes that media companies and authors have initiated against tech corporations because they use their material to train AI models without authorization. OpenAI repeatedly faces accusations of copyright infringement. The New York Times accuses the company of having unlawfully used articles from the newspaper for AI training. The parent company of PCMag and IGN, the media house Ziff Davis, also sued OpenAI for copyright infringement, as did several major Indian media outlets and a coalition of major Canadian media houses.
Other media companies, however, have decided to license their texts to OpenAI. For example, the German Axel Springer publishing house concluded a contract for a seven-figure sum, giving OpenAI access to texts from magazines such as BILD, Welt, or Politico.
(akn)
