Writing The Payout
Tech progress cannot come at the cost of creators’ rights
It’s a historic first. A US federal judge approved a $1.5bn settlement in a class-action suit against Anthropic, over its use of copyrighted books to train its AI model. The judgment is the first such settlement among the many lawsuits against Big Tech companies developing AI models, including OpenAI, Microsoft and Meta. For decades, companies have scraped original creative and scientific works off the internet without permission, let alone compensation. Only when the AI models came into everyday use, incorporated in a zillion ways in daily life, did the enormity of the data scraping hit us. Just last week, major book publishers, including Hachette and Elsevier, filed another class-action lawsuit against Google, alleging copyrighted books from Google Books were secretly repurposed to train Gemini AI.
The Anthropic settlement is one step closer to fixing accountability and signalling to all companies that there is a cost to overriding creators’ rights. While the court ruled that training AI on creative works can qualify as fair use, it found Anthropic acted unlawfully in storing millions of pirated books obtained from shadow libraries such as LibGen and Pirate Library Mirror. Around 500,000 works are covered; authors and publishers are expected to receive about $3,000 per work, shared according to ownership. That’s a relatively small amount, given the riches of AI companies.
While many AI companies now have licensing agreements with publishers, news organisations and content owners for use of material/content, the copyright issue hasn’t been addressed satisfactorily. To that end, the distinction this judgment made between what constitutes fair use and storing of pirated books, has not decided the core issue of copyright infringement. What constitutes infringement, what is fair use, is AI training rewarding piracy, these are as yet unresolved issues. The Anthropic settlement is not the final word, but it’s a significant beginning.
https://www.reuters.com/video/watch/idRW957021072026RP1/?chan=business
https://www.theguardian.com/books/2026/jul/14/publishers-sue-google-gemini-ai-training
Disclaimer
Views expressed above are the author’s own.
