The lawsuit also alleges that Anthropic used copyrighted material to train its language models without permission. The music industry has been grappling with copyright infringement issues in generative AI, with several lawsuits filed against platforms like ChatGPT, Stable Diffusion, and Midjourney. Despite these challenges, UMG announced its collaboration with companies like Google to guide its approach to generative AI. Anthropic, which received significant investments from Amazon and Google, insists it takes trust and safety seriously and trains its AI systems to follow a set of rules.
Key takeaways:
- Universal Music Group and other music publishers have sued AI company Anthropic for distributing copyrighted lyrics with its AI model, Claude 2.
- The complaint alleges that Claude 2 can generate almost identical lyrics to copyrighted songs and uses phrases extremely similar to existing lyrics, even when not asked to recreate songs.
- The plaintiffs accuse Anthropic of distributing copyrighted material without permission and using it to train its language models, which they consider as theft.
- Anthropic, which has received significant investment from companies like Amazon and Google, claims to take trust and safety seriously and bases its product and research principles on "constitutional AI".