Judge dismisses authors’ copyright lawsuit against Meta over AI training

1 comment

1. This is actually the second case decided this week on this issue –

   >On Monday, from the same courthouse, U.S. District Judge William Alsup [ruled that AI company Anthropic didn’t break the law](https://apnews.com/article/anthropic-ai-fair-use-copyright-pirated-libraries-1e5cece51c2e4bd0bb21d94de2abb035) by training its chatbot Claude on millions of copyrighted books, but the company must still go to trial for illicitly acquiring those books from pirate websites instead of buying

   To clarify what these rulings mean, *AI companies are free to trains their LLMs on copyrighted books, so long as they actually buy the books. NOT buy a LICENSE or RIGHTS to the books, just the books themselves.* They’ll be pretty happy with that result, as it’ll only cost them on the order of thousands to maybe millions dollars to purchase the material, compared to buying the rights to it. Meta said as much –

   >Meta said it appreciates the decision.

   The ruling is that training LLMs is ‘fair use’ as it counts as sufficiently ‘transformative’ –

   >the actual process of an AI system distilling from thousands of written works to be able to produce its own passages of text qualified as “fair use” under U.S. copyright law because it was “quintessentially transformative,” Alsup wrote

   This lends more backing to the idea that AI isn’t simply ‘copying’ or ‘stealing’, as critics often characterize it, but rather ‘training’ and ‘learning’ as AI supporters argue, akin to humans learning from the works of others. If that argument holds for books, it’s likely going to be the precedent for art and music as well…

Comments are closed.