In early November 2023, reggaetón fans excitedly listened to a slew of new Bad Bunny songs that flooded TikTok. While the vocals, lyrics, and music style appeared to belong to Bad Bunny, he did not write, record, produce, or much less release the viral hits. Instead, TikTok users used AI tools to create songs “in the style of” Bad Bunny.

Because of the frenzy around AI songs on TikTok and other platforms, many of which offer monetary compensation per view, major stakeholders in the music industry are worried about the copyright repercussions of AI tools. A senior VP from Universal Music Group (UMG) stated that generative AI infringes the label’s copyrights. But the law remains unsettled. Some claim that the model's training on copyrighted works without the artist's consent constitutes infringement, while AI developers claim this is fair use. And unless a rights-holder can prove that the AI-generated output is a copy or derivative of the artist's work, it is unclear whether AI-generated music “in the style of” a famous artist infringes that artist’s copyright.

Congress is seeking to tackle the unauthorized use of voices and visual likenesses in AI-generated content through laws akin to state-level right of publicity laws. In October 2023, a group of senators proposed the “NO FAKES ACT,” which seeks to protect those whose voice and likeness are mimicked via AI tools without permission. And just earlier this year, a bipartisan group of representatives proposed the “No AI FRAUD” Act,” also seeking to stop “unauthorized fakes using the images and voices of others.”

Can Creators of AI-Generated Music Copyright their “Creations”?

Although there is one notable instance where the U.S. Copyright Office granted copyright to a work with AI-generated elements, it said only the written and visual elements that the human “select[ed], coordinat[ed], and arrang[ed]” held the copyright. Federal courts have affirmed the Copyright Office’s guidance that AI-generated art does not qualify for copyrights because they are not authored by humans, highlighting that “[h]uman authorship is a bedrock requirement of copyright.”[1] Albeit, the Office noted it conducts a case-by-case analysis of “how the AI tool operates and how it was used to create the work.” Thus, whether an AI-generated song is copyrightable appears to be a case-by-case inquiry, analyzing whether there are any elements with human authorship, e.g., lyrics written or arranged by a human.

However, in August 2023, the Copyright Office published a “notice of inquiry and request for comments” on AI and copyright, specifically asking those submitting comments to address if “there [are] circumstances when a human using a generative AI system should be considered the ‘author’ of material produced by the system” as well as querying if the Copyright Act should be modified “to clarify the human authorship requirement or to provide additional standards to determine when content including AI-generated material is subject to copyright protection.”[2]

Can AI Tools and Music Coexist?

Some industry stakeholders have opted to embrace the use of AI for voice replication in music. Google and YouTube, both a part of Alphabet, recently partnered with nine artists to launch a music AI tool. The tool allows users to type a prompt, and the output will feature “the AI-generated voice” of one of the nine artists that the user preselects to use in a 60-second YouTube video. Separately, artists like Grimes have embraced the new technology, even releasing her AI tool. She is allowing users to commercially release music featuring an AI imitation of her voice through a 50-50% royalty split between her and the AI music creators. The AI tool developer and Grimes will “not claim any ownership of the sound recording or the underlying composition (unless that composition is a cover of a Grimes song).” Grimes has even gone so far as to declare that she “like[s] the idea of open sourcing all art and killing copyright.” Even UMG, which has been very vocal about its concerns about AI tools in the music space, partnered with YouTube for an incubator, allowing some of its artists to experiment with AI with a “commitment to… responsible innovation.” 

Despite efforts to regulate or embrace this new technology in the music space, copyright and other intellectual property laws have yet to meet the needs of this new technology. The landscape of AI voice replication remains unsettled.

 

[1] Thaler v. Perlmutter, No. CV 22-1564 (BAH), 2023 WL 5333236, at *4 (D.D.C. Aug. 18, 2023); see also U.S Copyright Office, Copyright Review Board, Second Request for Reconsideration for Refusal to Register SURYAST at 7, (Dec. 11, 2023), https://copyright.gov/rulings-filings/review-board/docs/SURYAST.pdf (affirming the Office’s refusal to register work created with AI because the work in question “is not the product of human authorship.”)

[2] 88 Fed. Reg. 59942, 59948 (Aug. 30, 2023), https://www.federalregister.gov/documents/2023/08/30/2023-18624/artificial-intelligence-and-copyright; compare with Thaler, 2023 WL 5333236, at *4 (holding that “[c]opyright has never stretched so far” as to protect new technologies that operate “absent any guiding human hand,” and that “[h]uman authorship is a bedrock requirement of copyright.”)