Anthropic can finally start writing checks to the group of authors and book publishers who sued AI Labs for copyright infringement. A federal judge on Monday gave final approval to a landmark $1.5 billion settlement in Anthropic’s copyright class action lawsuit, Reuters reports.
Judge William Alsup of the U.S. District Court for the Northern District of California granted preliminary approval to the settlement last year, ruling that Anthropic illegally downloaded and stored millions of copyrighted books.
Mr. Alsup has since retired, and Judge Araceli Martinez-Holguin signed the settlement on Monday.
The payments will include an estimated $3,000 per work for an estimated 500,000 works, to be divided between the authors and publishers who hold the rights to them. Although the settlement is considered the largest in U.S. copyright history, many writers and creators still do not consider it a victory.
It depends on how the legal issue was resolved. Alsup sided with Anthropic on core issues. He ruled that training AI models on copyrighted text qualifies as fair use. This decision is widely seen as a turning point for the AI industry. But the ruling did not excuse how Anthropic acquired the books in the first place. Anthropic was building its training library from two sources. One is the books you bought and scanned (which is fine), and the other is the books you downloaded from pirated sites like Library Genesis or Pirate Library Mirror. Alsup found the second method to be illegal on its own terms and said the copyright infringement issue could be taken to court. Anthropic quickly agreed to a settlement to avoid a trial and possible damages awards from a jury.
Although the final approval ends the case, it does not resolve the industry-wide legal problem, as the Alsup decision is a single district court decision and Anthropic’s settlement decision means the case will never go to an appellate court to become a legally binding precedent.
Other judges are still free to reach their own conclusions about the facts, and that is exactly what is happening elsewhere. Copyright lawsuits continue to mount against companies like Google, Meta, Midjourney, and OpenAI over whether it’s legal to train AI models with copyrighted works. Just last week, a group of publishers and authors, including Hachette, Cengage, Elsevier, author Scott Turow, and SCRIBE, filed a class action lawsuit against Google, accusing Google of using their copyrighted material to train its AI platform, Gemini.
If you buy through links in our articles, we may earn a small commission. This does not affect editorial independence.
