August 20, 2024 A group of authors has filed a lawsuit against Anthropic, an AI startup, accusing it of using pirated copies of their copyrighted books to train its chatbot, Claude. This marks the first legal action by authors against Anthropic, which has marketed itself as a responsible AI developer. The lawsuit, brought by authors Andrea Bartz, Charles Graeber, and Kirk Wallace Johnson, claims that Anthropic’s actions violate copyright laws and undermine its stated goals of ethical AI development.

The authors argue that Anthropic’s use of datasets containing pirated books for training its AI models constitutes “large-scale theft” of their intellectual property. They accuse the company of profiting from their creative work without permission or compensation, making a mockery of its claims to ethical AI practices.

The lawsuit also disputes the company’s defense that its actions fall under the “fair use” doctrine, highlighting that AI systems are not learning in the same way humans do, as they are consuming vast amounts of data without purchasing or obtaining the works. They allege that Anthropic trained its Claude chatbot on a dataset called “The Pile” that included a trove of pirated books, without permission or compensation to the authors.  This, according to the plaintiffs, constituted “illegal strip mining” of copyrighted content.

This case joins a growing list of lawsuits against AI developers for similar copyright infringement issues, with other notable cases involving OpenAI, Microsoft, and major media outlets. These legal challenges are part of a broader debate about the ethical and legal implications of AI training practices, especially as they relate to the use of copyrighted materials. As AI technology continues to advance, the resolution of these cases could have significant implications for the future of content creation and intellectual property rights.