12 hours ago
Anthropic wins AI training ruling but must answer for book theft
A federal judge ruled that
Anthropic
's use of copyrighted books to train its
Claude chatbot
constitutes fair use under copyright law, delivering a significant victory for the AI industry while ordering the company to face trial over acquiring pirated materials.
Tired of too many ads? go ad free now
US District Judge William Alsup of San Francisco determined that training AI models on copyrighted works was "quintessentially transformative" and legally justified. The ruling dismissed key copyright infringement claims brought by authors Andrea Bartz, Charles Graeber, and Kirk Wallace Johnson, who sued Anthropic last year alleging "large-scale theft" of their works.
The court separated how AI companies use books from how they obtain them
Alsup's decision hinged on the
transformative nature of AI
training, comparing it to "any reader aspiring to be a writer" who learns from existing works to create something entirely different.
The judge emphasized that Anthropic's large language models didn't reproduce the authors' creative elements or identifiable writing styles, making the training process legally permissible under fair use doctrine.
However, the company isn't completely off the hook. Alsup ordered Anthropic to face a December trial over allegations that it illegally downloaded millions of books from online "shadow libraries" of pirated copies.
Court documents revealed internal employee concerns about using pirate sites before the company shifted strategies.
Court's ruling could influence similar lawsuits against OpenAI and Meta
Anthropic later hired former Google Books executive Tom Turvey and began purchasing books in bulk, physically scanning them for AI training. But the judge noted that buying legitimate copies after initially using pirated versions won't absolve the company of potential liability for the earlier theft, though it may reduce statutory damages.
The decision establishes important legal boundaries as AI companies face mounting copyright challenges, potentially setting precedent for similar cases against ChatGPT-maker OpenAI and Meta.
