Report from Engadet
In Brief – Federal District Judge Araceli Martinez-Olguin has approved Anthropic’s record $1.5 billion settlement of a copyright lawsuit brought by authors who accused the company of violating their copyrights to train its Claude AI models. The case’s original judge, William Alsup, ruled that training Claude with copyrighted works qualified as fair use but found that Anthropic infringed authors’ copyrights by storing seven million pirated books in a central library. Although Alsup, who has since retired, declined to approve an earlier version of the settlement, Martinez-Olguin approved revised terms requiring Anthropic to pay authors of more than 480,000 works $3000 per eligible book, as well as destroy the pirated copies.
Context – The issue of applying copyright law to AI training is being debated globally, but major copyright lawsuits in US courts will be the most impactful. The key question remains the application of the “fair use” doctrine. The Trump Administration’s AI legislative framework says that it believes that model training is fair use and that Congress should allow the federal judicial branch to resolve the issue. Two conflicting court opinions released last summer highlight the complexities. Judge Alsup’s vigorous defense of generative AI training as “fair use” was countered the following week by Judge Vincent Chhabria who dismissed a copyright lawsuit targeting Meta’s AI training, but in doing so argued that AI training was not fair use. He created the novel concept of “indirect substitution” through which AI systems nullify the fair use defense by creating massive volumes of cheap content that did not technically copy originals but were “similar enough to compete with the originals and thereby indirectly substitute for them”. Creative industries have strongly criticized proposals in the UK and Australia to make their copyright regimes more conducive to AI training, as well as Japan’s existing training-friendly regime. An Indian panel rejected applying fair use for training and proposed mandatory royalty payments for AI training by tech giants that would be deposited in a centralized licensing pool that all Indian creators would draw from.