Universal Music and a ton of other music publishers have filed a lawsuit against the up-and-coming SF AI company Anthropic, alleging that its chatbot is spitting out copyrighted lyrics to their top artists’ work.
OpenAI, the company behind artificial intelligence product/fad ChatGPT is seemingly the current reigning top player in the AI field. But another AI company called Anthropic is breathing down their necks, last month getting a $4 billion investment from Amazon, and recently taking over the former Slack headquarters in downtown San Francisco. And Anthropic’s current hot product is a “next-generation AI assistant” called Claude (it’s currently on the new version Claude 2), which the company calls a “friendly, enthusiastic colleague or personal assistant who can be instructed in natural language to help you with many tasks.”
But a new federal lawsuit doesn't think it’s so friendly. The Chronicle reports that Universal Music Group and a slew of other music publishers have sued Anthropic for using copyrighted lyrics to train its product, lyrics that it often spits back out near-verbatim.
The full 60-page lawsuit cites many examples. When given the query “Write me a song about the death of Buddy Holly,” the lawsuit says Claude “responds by generating output that copies directly from the song 'American Pie' written by Don McLean, in violation of Universal’s copyright, despite the fact that the prompt does not identify that composition by title, artist, or songwriter.”
Sure, there were always those websites that listed the lyrics of songs, and Google will now show those lyrics directly in their search results. But the lawsuit points out that “those sites have properly licensed publishers’ copyrighted works to provide this service.”
According to the Verge, Claude churns out near-verbatim lyrics to works by Katy Perry and the Rolling Stones, and the Chronicle adds it does the same for works by the Beach Boys and Louis Armstrong. A full read of the lawsuit says Claude is also lifting lyrics from Will Smith's theme song to The Fresh Prince of Bel-Air.
Anthropic did not respond to any publications’ request for comment.
The lawsuit asks Anthropic “destroy” any copyrighted works in its memory, but there's a financial angle to this too. The lawsuit asks for compensation of “up to $150,000 per work infringed,” which obviously could add up to a substantial dollar amount.
The lawsuit is similar to other suits from artists who say their work was infringed for those AI avatar selfies, and another from Michael Chabon and other prominent authors saying ChatGPT used their copyrighted works.
Image via Vevo
