ANTHROPIC
Az Anthropic győzelmet aratott az AI „fair use” igényével kapcsolatban
A federal judge just ruled that Anthropic's AI training on legally purchased books qualifies as fair use, but rejected any defense for 7M pirated copies in its digital library in the company’s class action lawsuit filed by a group of authors.
- The judge called AI training "spectacularly" transformative, comparing Claude to aspiring writers learning from established authors rather than copying them.
- The authors failed to demonstrate that Claude could generate outputs resembling their original works, weakening core claims about competitive harm.
- The filings revealed that Anthropic legally spent “many millions” to purchase print books, scanning them into digital files for use in AI training.
- However, Anthropic also downloaded millions of books from pirate sites, storing them permanently, which the court said violated authors’ rights.
- The company will face trial in December for willful infringement of the pirated works, with potential damages potentially reaching $150,000 per book.
Miért fontos?
AI labs get a green light to train on legally obtained data, which could provide one of the first precedents for countless cases currently levied against tech companies. However, given the lack of clarity around how many copyrighted materials get into training data, this is still just one battle in what looks like a very long legal war.