Two cases in the Northern District of California have had oral arguments on the respective defendant’s motion for summary judgment on fair use: Kadrey v. Meta and Bartz v. Anthropic.
Both involve lawsuits brought by book authors. Both involve the defendant’s use of so-called pirated books datasets, or shadow libraries, which defendants contend were used to get copies for the datasets they used to train their respective AI models, LlaMa and Claude. Both cases have been fully briefed and argued–with pretty much the same strategy: defendants seek summary judgment on fair use but plaintiffs argue that the defense contains factual disputes that require a trial.
And, in both cases, Judge Chhabria and Judge Alsup, respectively, voiced concern on how to weigh the AI company’s use of pirated books datasets, or shadow libraries offered online without authorization of the books’ authors. Yet, both judges were leaning to recognizing the defendant’s use of the copies of works was for the transformative purpose to develop and train AI, a new technology. Judge Chhabria even suggested during the hearing that it might be a “highly transformative” purpose.
Judge Alsup even went so far as to tell the parties that he’s leaning to finding a fair use (but could change his mind). Judge Chhabria also voiced concern about the potential for AI to “obliterate” the market with millions of works and thereby destroy the value of works especially of new authors, such as the next Taylor Swift. But he also questioned whether the Kadrey plaintiffs had even raised this theory and presented sufficient evidence to raise a genuine issue of material fact. By contrast, Judge Alsup expressed greater skepticism to the Bartz plaintiffs’ similar argument of market harm based on the proliferation of AI-generated works–calling it “baloney” and “science fiction.”
The hearing in Kadrey was on May 1. So Judge Chhabria does have a 3 week head start on Judge Alsup. It would be nice if the opinions are published on the same day!
Related stories