Artificial intelligence (AI) is now part of daily life, powering customer service chatbots, virtual assistants like Siri and Alexa, automated email responses, and personalized shopping recommendations. But as these systems get smarter, they need ever-larger amounts of data to learn, often drawing on copyrighted books and creative works. This has led to new legal battles over whether AI companies are crossing the line into copyright infringement, or whether their use of these materials to train large language models qualifies as “fair use.”
Two recent cases, Kadrey v. Meta Platforms, Inc. and Bartz v. Anthropic PBC have brought significant attention to the legal complexities surrounding the use of copyrighted works in AI systems.
In this client alert from the Vorys intellectual property group, we dive into the details of these cases.
If you are an author or creator, these cases show that your work might be used to train AI models without your knowledge, but they also demonstrate that legal remedies are available, especially if your work was pirated and used by a tech company. For businesses and startups, the lesson is clear: use only properly licensed training data, as relying on pirated or unlicensed material can result in expensive lawsuits and reputational harm. Consumers should expect more legal battles over who owns the content they see online and how it was created, since some AI-generated text, images, or music may be based on copyrighted works, raising important questions about originality and ownership.
As the law catches up to the rapid growth of AI, courts are beginning to set boundaries for how creative works can be used to power new technology. The outcomes of these cases, and settlements like Anthropic’s, will help define what’s fair, what’s legal, and what’s possible in the age of AI.
Client Alert Authors:
William Oldach III and Athena Williams