In a significant development for the field of artificial intelligence and copyright law, a recent ruling in the US District Court for the Northern District marked a pivotal moment for the fair use doctrine as it applies to generative AI. Judge William Alsup’s decision in favor of Anthropic PBC, a company associated with Amazon, has provided a framework for handling generative AI cases that are currently making their way through the courts.
This decision affirms that using legitimately-acquired materials to train generative AI models is fair use under US copyright law. Specifically, Judge Alsup ruled that Anthropic’s use of “Claude,” their large language model, in training activities constitutes transformative fair use. This means that purchasing a book, training the AI model with its content, and subsequently destroying the book is an acceptable practice.
However, the ruling was not an unequivocal win for Anthropic. Claims related to the use of works sourced from pirated book websites were not summarily dismissed. Instead, Judge Alsup has allowed these allegations of copyright infringement to proceed to trial. This indicates that while the ruling supports the fair use defense in specific contexts, it also underscores the ongoing legal challenges related to unauthorized sources.
The court’s approach in this mixed ruling is likely to serve as a reference for judges handling similar cases involving generative AI and copyright claims. The decision could have far-reaching implications, offering guidance to tech companies and legal professionals about the bounds of fair use in the rapidly evolving AI sector.
For more detailed insights into the ruling and its implications, you can read the full article on Bloomberg Law.