The Value of IP in an Emerging AI World
by Michael Antone, Counsel
Anthropic PBC recently agreed to a $1.5 billion settlement in the class action lawsuit brought by publishers and authors, Bartz v. Anthropic, USDC CA-ND, Case 3:24-cv-05417-WHA (“Bartz”), for damages related to the use of copyrighted works for AI training. Authors and publishers that believe that their works may be involved in the settlement should visit the Anthropic Copyright Settlement Website for more information.
While the Bartz settlement is a landmark, it is really only the tip of the iceberg regarding copyright infringement claims against AI companies that have used copyrighted works to train AI models without permission from the copyright owners.
The settlement was limited to those copyrighted works that Anthropic used retrieved from unlicensed (pirate) library sites like LibGen and the Pirate Library Mirror (“PiLiMi”) in developing its Claude large language model (LLM). The settlement followed a June 2025, summary judgement ruling that Anthropic’s keeping of pirated copies in its internal “central library” was not protected under copyright law.
The ruling in Bartz left open the bigger question of whether the use of legally obtained copies of copyrighted works to train AI models is fair use. This is a high stakes question in the AI area and there are strong and compelling views on both sides of the issue.
Whatever side you take on the fair use of protected materials by AI models, your intellectual property rights will be impacted by AI and the impact will be swift and sweeping.
Read more in our Copyrights FAQ or Contact Us to speak with an attorney about your IP concerns.
