According to reports, the artificial intelligence startup Anthropic is facing a class-action lawsuit in the Federal Court of California.
According to IntelliNews, the artificial intelligence startup Anthropic is facing a class-action lawsuit in the Federal Court of California. Three authors have accused the San Francisco-based company of misusing the works of hundreds of thousands of people, including themselves, to train the AI chatbot Claude.
On Monday, authors and journalists Andrea Bartz, Charles Graeber, and Kirk Wallace Johnson filed a lawsuit, claiming, "Anthropic has built a multi-billion dollar business by stealing hundreds of thousands of copyrighted books. Anthropic did not obtain permission or pay a fair price for the works it used, but engaged in piracy."
The plaintiffs claim that their works and those of others were used to teach the company's large language model (LLM) Claude. They claim, "An important component of Anthropic's business model, as well as its flagship 'Claude' series of large language models, is the large-scale theft of copyrighted works."
Following last year's accusation by music publishers of the misuse of copyrighted lyrics to train Claude, this lawsuit is the second against the company. Anthropic is backed by Amazon and Google-owned Alphabet. Currently, antitrust agencies in the United Kingdom and the United States are investigating Anthropic's collaboration with Google and Amazon.
The plaintiffs claim that despite Anthropic keeping the sources of their training corpus secret, they further add that one of the creators of The Pile is an independent developer named Shawn Presser, who created a dataset called "Books3" which contains a large number of pirated books.
The plaintiffs are seeking an unspecified amount of monetary compensation and requesting the court to permanently enjoin Anthropic from engaging in the alleged copyright infringement.
This lawsuit is one of several legal actions faced by companies developing artificial intelligence models for alleged misuse of copyrighted materials. Different groups of authors have filed lawsuits against OpenAI and Meta, supported by Microsoft (MSFT.US), accusing them of abusing their work to train large language models that power chatbots.