Irving Azoff’s Music Artists Coalition (MAC) has joined a brief in a United States (US) copyright case brought by the New York Times against OpenAI and Microsoft. The outcome could shape how courts treat the use of copyrighted works to train artificial intelligence (AI) models.
Coalition and music groups file briefs#
MAC signed on alongside the Human Consent Foundation (HCF), a nonprofit co-founded by actor Cate Blanchett, and the talent agency CAA. On October 30, the group asked the court for permission to submit a brief on whether rights to creative works can be cleared and licensed for AI use at scale.
The same day, the Recording Industry Association of America (RIAA), the National Music Publishers’ Association (NMPA), the American Society of Composers, Authors and Publishers (ASCAP), Broadcast Music, Inc. (BMI) and five other music organizations filed a proposed brief of their own. That filing formally backs the plaintiffs.
The wider case and the fair use question#
The case combines the Times’ lawsuit with others against OpenAI and Microsoft, including suits brought by book authors. Both sides have asked Judge Sidney Stein to settle its central question without a trial: whether training AI models on copyrighted works without a license is “fair use.”
Fair use is an exception in US copyright law that allows copyrighted material to be reused without permission in certain circumstances. Microsoft has asked the court to hold that training a general-purpose large language model (LLM) on copyrighted works is fair use as a matter of law, according to the RIAA brief.
The music groups are not parties to the case, and their briefs do not seek damages from OpenAI or Microsoft. A ruling in the tech companies’ favor could, however, give AI firms facing music-industry lawsuits, including Suno, Udio and Anthropic, a fair use decision from a New York federal court to cite in their own defense. Sony Music has two lawsuits against Udio pending in the same court.
The HCF-led brief#
The HCF filing says the requests from OpenAI and Microsoft raise questions over whether rights for AI training “can practically be identified, cleared, and licensed at scale, and whether a functioning market exists for those uses.”
HCF publishes a machine-readable standard that lets creators and rightsholders declare how AI systems may use their work and likeness, according to the filing. The group says it draws on experience in “creative-rights licensing, representation, and machine-readable consent infrastructure.” It notes that its brief takes “no position on the ultimate disposition” of the motions before Stein.
The brief argues that the creative industries have “a proven infrastructure and workflows to clear rights for dispersed works at scale.” It answers the tech companies’ “dispersed owner” concern by pointing to collective licensing, citing the International Confederation of Societies of Authors and Composers (CISAC) and BMI as examples. Its appendix lists supporters including Madonna, Bette Midler, Anne Hathaway and Emma Thompson.
The other signatories are the National Association of Voice Actors, enterprise AI company Writer, the Creators Coalition on Artificial Intelligence, Blanchett, filmmaker Steven Soderbergh and 311 frontman Nick Hexum. Soderbergh is a founding board member of HCF, and Hexum, like Blanchett, is a co-author of its consent standard.
The filing describes MAC, quoting the group’s website, as focused on “protecting the rights, financial security, and creative freedoms of musicians.”
The RIAA-led brief#
The brief from the RIAA, the NMPA and their co-signers asks Stein to rule that OpenAI and Microsoft’s training on the plaintiffs’ works is not fair use. The co-signers are the American Association of Independent Music (A2IM), Songwriters of North America, the Black Music Action Coalition, the Recording Academy, ASCAP, BMI and the Artist Rights Alliance. The nine groups claim their members “own or control the majority of music consumed in the United States.”
“AI companies copy millions of copyrighted sound recordings and musical compositions without authorization to train their models to generate a limitless supply of synthetic substitutes that invade the music marketplace. AI-generated music floods streaming platforms, tops charts, and dominates social media playlists. It displaces human-authored music and deceives consumers, who cannot tell whether they are listening to human artists or to outputs created by a computer,” the RIAA brief states.
Purpose of copying and training#
The brief argues that OpenAI and Microsoft describe the purpose of their copying too broadly, as teaching AI to communicate and “spread knowledge.”
“An AI model does not ingest romance poetry or bluegrass to learn about language or music at large, but rather to generate substitutes that read and sound like them in practice,” the brief tells Stein.
It adds that the purpose of training on “a John Grisham novel or a Stevie Wonder song” is to teach a model to produce “a Grisham-like legal thriller or a Wonder-like audio file.”
The US Justice Department has compared AI training to author Joan Didion drawing on Ernest Hemingway as an influence, the brief says.
“Law-abiding authors pay for the modest universe of works that inspire them,” the RIAA and its co-signers respond. “They do not steal them by the millions and encode copies on their brains like an AI system does.”
Licensing arguments and deals cited#
On licensing, the RIAA brief contends: “It is thus a myth that licensing copyrighted works for AI training is either too ‘sporadic’ or speculative to establish a ‘reasonable licensing’ market.” It rejects Microsoft’s argument that the authors whose works it would need to license are “dispersed, difficult to identify, and costly to negotiate or contract with.”
The brief points to the following AI licensing deals:
- All three major labels with KLAY
- Udio with Universal Music Group, Warner Music Group and Merlin
- Suno with Warner, BMG and Believe
- ElevenLabs with Universal, Kobalt and Merlin, among others
“The music industry is as fragmented as the book and news publishing industries, and yet copyright owners of sound recordings license nearly all their recorded music to streaming platforms,” the brief states.
