The epic class action lawsuit concerned a library of 7 million pirated books used in training, and had Anthropic looking at $150,000 in penalties per instance of copyright theft, but it was settled last week without disclosing terms.
Largest copyright settlement ever
Now that the terms are clear, it is still enough to sting the AI giant — paying out $3,000 per book for an estimated 500,000 books in the database, totaling 1.5 billion dollars.
It could reach even higher if new claims emerge.
— If approved, this landmark settlement will be the largest publicly reported copyright recovery in history, larger than any other copyright class action settlement or any individual copyright case litigated to final judgment, said Justin Nelson, one of the author’s attorneys, according to NBC News.
Open for claims
Lawyers for the authors have put up the webpage anthropiccopyrightsettlement.com to support claimants for the huge database of books, where «potential class members,» as in authors of copied books, can stake their claim, writes The Verge.
The lawsuit will now go onto a hearing on the settlement on September 8th, where it must be approved by the court before going into effect.
— Today’s settlement, if approved, will resolve the plaintiffs’ remaining legacy claims. We remain committed to developing safe AI systems that help people and organizations extend their capabilities, advance scientific discovery, and solve complex problems. Anthropic said in a statement to many outlets.
«Fair use» argument failed on pirated books
Anthropic had originally argued use of the books was «fair use,» insofar as they were being used to create a «transformative» work, not in competition with the copyright holders.
Initially, the court of US District Judge William Alsup agreed, but only on purchased books, calling it little different from teaching kids how to read.
Alsup did, however, hold that Anthropic’s library of 7 million pirated books was a possible copyright violation and agreed to run it as a class action suit.
Possible ramifications
— [This is] a vital step in acknowledging that AI companies cannot simply steal authors’ creative work to build their AI, said Mary Rasenberger, CEO of the Authors Guild, on the settlement, to Reuters.
There are many other court cases on the same theme against several AI labs yet to be resolved, but this is the first, and maybe the largest one to have gone to fruition.
— This settlement will push other AI companies to the negotiating table and accelerate the creation of a true marketplace for data, likely involving API authentications and revenue-sharing models, opines Cecilia Ziniti, founder and CEO of AI legal company GC AI, to Deadline.com.
Read more: The actual settlement, writeups on NBC News, The Verge, Reuters and Deadline.com. Discussion on r/Singularity.