SAN FRANCISCO: A United States judge has granted preliminary approval to a settlement requiring artificial intelligence company Anthropic to pay $1.5 billion to resolve a class action lawsuit.
The lawsuit accused Anthropic of building a library of pirated books to train its Claude AI chatbot.
Anthropic deputy general counsel Aparna Sridhar expressed satisfaction with the court’s decision to grant preliminary approval.
She stated that the ruling allows the company to concentrate on developing safe AI systems.
The legal action originated from a complaint filed by authors Andrea Bartz, Charles Graeber, and Kirk Wallace Johnson.
These authors alleged that Anthropic illegally copied their books to train its Claude AI system.
In a significant June ruling, US District Court Judge William Alsup determined that training AI models with books constituted transformative fair use.
Judge Alsup compared the AI training process to human learning through reading.
However, the judge rejected Anthropic’s argument that downloading millions of pirated books for a permanent digital library was protected under fair use.
Anthropic’s legal counsel emphasised that the settlement resolves specific claims about material acquisition methods while preserving the core fair use ruling.
The settlement encompasses approximately 500,000 books, amounting to roughly $3,000 per work.
This per-work payment is four times the minimum statutory damages available under United States copyright law.
As part of the agreement, Anthropic must destroy the original pirated files and any copies it created.
The company retains the right to use books it legally purchased and scanned for its operations.
Authors Guild chief executive Mary Rasenberger stated that the settlement delivers a strong warning to the AI industry about the consequences of pirating authors’ works.
She highlighted the impact on authors who are least able to afford such exploitation.
Anthropic, based in San Francisco, recently secured $13 billion in funding, valuing the AI startup at $183 billion.
The company competes with generative AI offerings from major tech firms like Google, OpenAI, Meta, and Microsoft.
This competition has created an immense demand for data, including books, to train increasingly sophisticated AI models. – AFP