(Reuters) – Official intelligence company Human Humans convinced California federal judges to reject preliminary bids on Tuesday to stop it from using lyrics owned by General Music Group and other music publishers to train its AI-powered Chatbot Claude.
U.S. District Court Judge Eumi Lee said the publishers’ demands were too broad and that their failure to show anthropomorphic behavior made them “irreparable.”
The publisher said in a statement that they “have more confidence in our cases against humanity.” A spokesperson for the personification said the company was pleased that the court did not approve the publisher’s “destructive and amorphous request.”
Music publishers UMG, Concord and Abkco sued humans in 2023, accusing it of infringing their copyrights, including musicians including Beyoncé, the Rolling Stones and the Beach Boys, lyrics for at least 500 songs.
The publisher claims that anthropomorphism uses lyrics without permission to train Claude to respond to human prompts.
The lawsuit is one of several debates, namely, the authors, news media, visual artists and others’ versions of works have been abused without consent or payment to develop AI products.
Technology companies, including OpenAI, Microsoft and Meta platforms, say their systems “make rational use” copyrighted materials under U.S. copyright law, learning to create new, transformative content by studying it.
Although Lee’s opinion did not specifically address the issue, fair use could be a decisive issue in the lawsuit.
Lee rejects the publisher’s argument that human use of its lyrics can’t be compensated by reducing the licensing market.
“Publishers essentially require courts to define the outline of the licensing market for AI training, where the issue of fair use thresholds remains unresolved,” Lee said.
(Report from Blake Brittain in Washington; Edited by Stephen Coates)
