Anthropic sued by US authors over use of 'pirated books' to train AI chatbot

The Authors’ Licensing and Collecting Society (ALCS) has described Artificial Intelligence (AI) company Anthropic’s alleged use of "pirated books" to train AI chatbot Claude as "egregious" and “typical of a wider trend.” 

Earlier this week, lawyers acting on behalf of authors Andrea Bartz, Charles Graeber, and Kirk Wallace Johnson filed a copyright infringement case in California, US, claiming Anthropic had used “pirated” copies of their books to teach its AI chatbot, Claude. 

The legal action accuses Anthropic of “downloading and copying hundreds of thousands of copyrighted books taken from pirated and illegal websites”.

This is the latest in a series of legal cases where AI firms stand accused of using authors’ material to teach large language models (LLMs) without their consent, or by using allegedly stolen copies of their books.

The case against Anthropic alleges the tech company committed “brazen infringement” of copyright, by downloading “known pirated versions” of the authors’ works, which it “fed” into its models. 

Documents for the court filing allege that Anthropic took “these drastic steps” to help its computer algorithms “generate human-like text responses”.

The three US-based authors are seeking consent and compensation as well as damages “for the large-scale infringement of copyrighted works” and “injunctive relief to prevent such improper conduct from ongoing and recurring”. 

Anthropic’s latest accounts suggest it will generate more than $850m of revenue in 2024. It has raised $7.6bn in funding from financial backers including tech giants Amazon ($4bn) and Google ($2bn).  

Anthropic is an offshoot of OpenAI, creator of ChatGPT. The latest iteration of Claude only launched 16 weeks ago, but yesterday it hit $1m in gross mobile app revenue across iOS and Android according to a trading update.

An Anthropic spokesperson told The Bookseller: “We are aware of the suit and are assessing the complaint. We cannot comment further on pending litigation.”

ALCS chief executive Barbara Hayes, said: “Sadly, this is typical of a wider trend whereby AI companies develop their models through the use of authors’ works without permission, recognition or remuneration. At the same time, those authors whose works are being infringed are seeing their incomes plummet. In the case of Anthropic, it appears even more egregious as the training data allegedly includes pirated content."

ALCS recently surveyed its members, receiving "an overwhelming number of responses from concerned writers", to explore licensing models that would give authors control over "how their works are used by AI companies and allow them to be remunerated for such uses, if that is what they choose".

In July, academics hit out at Taylor & Francis (T&F) for selling access to its authors’ research as part of a partnership with Microsoft worth $10m, with parent firm Informa’s half-year financial results later revealing that it was set to earn £58m ($75m) from selling access to its authors’ work to AI firms. Two further academic publishers—Wiley and Oxford University Press—subsequently confirmed they have made deals with or are considering working with artificial intelligence (AI) companies, with Wiley revealing in its latest trading update that it earned $23m from giving an unnamed company access to its content to train its LLMs.

The Society of Authors’ (SoA) senior policy and public affairs manager, Ambre Morvan, told The Bookseller: "The Society of Authors is very concerned about the danger generative artificial intelligence (AI) poses to authors’ livelihoods and to the future of the profession, and the potential damaging implications of people reading and relying on AI-interpreted output. Numerous cases have now been reported, in the UK and overseas, of these systems being ‘trained’ and developed on datasets which include a significant number of copyright protected works. We are closely monitoring these cases which, in our view, demonstrate the urgency of the situation and need for immediate action."