A recent decision in the U.S. has set a powerful precedent for creators worldwide: using copyrighted content to train AI models does not automatically qualify as fair use. In the dispute between Thomson Reuters and Ross Intelligence the court ruled that Ross infringed Westlaw’s copyrights by feeding its legal database into an AI tool meant to compete with Westlaw’s services. While the case focused on a non-generative AI application, it sends a clear signal to all industries about the importance of respecting copyright when training artificial intelligence.

At Copyright Power International we see AI as a transformative force in music production and rights management, but transformation must happen responsibly. Models like Suno and Udio that ingest copyrighted music to learn melodies and styles now face legal scrutiny similar to the one Ross encountered. This ruling underscores the need for transparent licensing frameworks and well-defined guidelines to ensure AI innovation does not trample artist derechos. We welcome Buma/Stemra’s recent clarification on what counts as copyrightable when AI is involved, as it offers a roadmap for fair practice.

Looking ahead, the music industry must work collaboratively. Publishers, labels, technology platforms and rights societies should craft regulations that balance creative opportunities with creator protections. Possible next steps include developing standardized AI training licenses, defining what constitutes transformative use in music, and establishing oversight bodies to monitor compliance. Clear rules will help AI tools thrive without eroding the livelihoods of songwriters, composers and performers.