Part 2 of 3 in our PlaiPals series.
This post is general information, not legal advice. For your own deals, talk to a music lawyer.
In Part 1, we saw how AI learned to copy singers before anyone asked them. The tools got good, the rules didn't exist yet, and famous voices became free raw material.
Here's the good news. Three groups have started pulling in the same direction. Artists built their own approved voice models. Labels started licensing instead of only suing. And lawmakers started writing the voice itself into the law.
First, one idea that makes everything else click.
Three things live inside every vocal track
Every sung record stacks three separate things together.
1. The song. The melody and lyrics. That's the composition, usually owned by the songwriters and their publishers.
2. The recording. The actual audio file of that performance. That's the master, often owned by a label or by an independent artist who kept their rights.
3. The voice. The sound that makes a singer recognizable as them. That belongs to the person.
Songs and recordings are protected by copyright. Voices, it turns out, mostly are not. In a July 2024 report, the U.S. Copyright Office concluded that a voice on its own isn't protected by copyright, even though a recording of that voice can be. The Office said existing laws weren't enough to deal with AI digital replicas and recommended that Congress create a new federal right (U.S. Copyright Office; full report).
Until then, voices have been protected mainly by state "right of publicity" laws, the same kind of law Bette Midler relied on when an ad agency hired a soundalike in the 1980s. The court in her case put it bluntly: "A voice is not copyrightable. The sounds are not 'fixed.' What is put forward as protectible here is more personal than any work of authorship" (Justia).
Why this matters: owning a song, or even a master, doesn't automatically give you the right to make brand-new vocals in the singer's voice. As the Los Angeles Times noted in September 2026, every contract is different, and it's often unclear how AI products will handle music where rights are split (Los Angeles Times). A voice is its own right. Treat it that way.
Artists who went first
Some artists didn't wait. They built their own.
Holly Herndon's Holly+ (2021). In July 2021, the experimental musician launched Holly+, a "digital twin" of her voice built with Never Before Heard Sounds. People could upload audio and get it back sung in her voice. The project's intellectual property was placed in a DAO, a kind of member-run online cooperative, with licensing proceeds going toward new development (The FADER). Herndon described the goal as a way to "allow people to perform through my voice, reduce confusion by establishing official approval, and invite everyone to benefit from the proceeds generated from its use" (The FADER).
Grimes and Elf.Tech (2023). In May 2023, Grimes launched Elf.Tech, software that turns your recorded vocals into her voice. Creators could release the results commercially in exchange for half of the master-recording royalties. She asked people to "be tasteful" and said she would block extreme uses (Pitchfork). Distributor TuneCore set up a process for releasing GrimesAI collaborations (TuneCore).
Kits.ai and paid voice models. Kits.ai runs a program called Kits Earn where singers train a verified AI model of their own voice and get paid when users download audio made with it (Kits.ai).
Different approaches, same idea: the singer says yes, is named, and shares in the money.
Labels move from lawsuits to licenses
In 2024, the major labels sued two AI song apps (see Part 1). In late 2025, several of those fights turned into deals.
- Universal and one AI song app (October 29, 2025). Universal Music Group settled its lawsuit with that app and signed licenses covering recorded music and publishing. The two announced a new subscription platform, trained on licensed music, planned for 2026. In the meantime, that app said creations would be kept inside a "walled garden" (UMG press release).
- Warner and that same app (November 19, 2025). Warner Music Group settled with the same AI song app. The planned service would let users make remixes, covers and new songs using the voices of artists and songwriters who choose to participate, "while ensuring artists and songwriters are credited and paid" (WMG).
- Warner and a second AI song app (November 25, 2025). Warner also settled with a second AI song app. Warner says its artists and songwriters get an opt-in for the use of their name, image, likeness, voice and compositions in new AI songs (WMG). That app also said it would launch licensed models in 2026 and require a paid account to download songs (BBC).
Where things stand as of late September 2026, based on the Los Angeles Times: the second app has released its first models built with licensed music from Warner, BMG and Believe. The first app has licensing deals with Universal and Warner, plus Merlin (which represents independent labels) and the publisher Kobalt. Sony and Universal are still suing the second app, and Sony has filed a second lawsuit against the first app (Los Angeles Times).
The same report adds a reality check: none of these companies have disclosed what artists will be paid or how many have opted in. The musicians' union AFM has sued Universal and Warner, alleging recordings its members played on were licensed to AI companies without paying those musicians; the labels have moved to dismiss (Los Angeles Times). One industry advisor quoted in the piece summed it up: "A label making a deal and an individual artist having a choice are not necessarily the same thing."
That gap is what we care about most. More in Part 3.
The laws: voice becomes a named right
Tennessee's ELVIS Act (2024). On March 21, 2024, Tennessee Gov. Bill Lee signed the Ensuring Likeness, Voice, and Image Security Act. It updated the state's personal rights law to add "voice" to what it protects, and the governor's office called it first-of-its-kind protection for musicians against AI misuse (Tennessee Governor's Office). It took effect July 1, 2024 (bill text).
California's digital-replica laws (2024). On September 17, 2024, Gov. Gavin Newsom signed two bills (Governor of California):
- AB 2602 requires contracts to spell out the use of AI-generated digital replicas of a performer's voice or likeness, and the performer must have professional representation when negotiating it (bill status).
- AB 1836 bars commercial use of digital replicas of deceased performers, including in sound recordings, without consent from their estates (bill status).
In plain words: in California, a vague contract line isn't supposed to be enough to sign away your digital voice, and a performer's voice doesn't become free to use just because they've died.
The federal NO FAKES Act (not law yet). The NO FAKES Act of 2026 (S. 4591) was introduced in the Senate on May 20, 2026, with a bipartisan group of sponsors. It aims "to protect intellectual property rights in the voice and visual likeness of individuals." The Senate Judiciary Committee advanced it in June, and on June 24, 2026, it was placed on the Senate calendar (Congress.gov). Earlier versions of the bill, first introduced in 2024, stalled in committee (Digital Music News). As of this writing, it has not passed the full Senate or the House. It is not law, and details could change.
What this means for you
If you're a singer: your voice is increasingly recognized as yours. Read any contract that mentions "digital replicas," "AI," "voice models" or "likeness" carefully. Opt-in should mean opt-in.
If you're a creator: using a real singer's voice without permission is getting riskier. The safe path is the one it always was: work with singers who said yes.
If you own masters: owning the recording isn't automatically a license to generate new vocals in the singer's voice. The fair move is to ask the artist.
The direction is clear: consent, credit and pay are becoming the price of admission. In Part 3, we'll show how we plan to build that into PLAIGROUND from day one.
Laws and deals described here were checked against the sources below as of September 2026 and may change.
Sources
- U.S. Copyright Office, NewsNet Issue 1048 (July 31, 2024): https://www.copyright.gov/newsnet/2024/1048.html
- U.S. Copyright Office, "Copyright and Artificial Intelligence, Part 1: Digital Replicas" (July 2024): https://www.copyright.gov/ai/Copyright-and-Artificial-Intelligence-Part-1-Digital-Replicas-Report.pdf
- Midler v. Ford Motor Co., 849 F.2d 460 (9th Cir. 1988), via Justia: https://law.justia.com/cases/federal/appellate-courts/F2/849/460/37485/
- Los Angeles Times, "Why some musicians aren't happy about labels signing AI deals with platforms" (Sept. 21, 2026): https://www.latimes.com/entertainment-arts/business/story/2026-09-21/ai-music-licensing-deals-labels-artists-who-benefits
- The FADER, "Holly Herndon announces new voice instrument and 'digital twin' Holly+" (July 14, 2021): https://www.thefader.com/2021/07/14/holly-herndon-holly-plus
- Pitchfork, "Grimes Unveils Software to Mimic Her Voice, Offering 50-50 Royalties for Commercial Use" (May 2, 2023): https://pitchfork.com/news/grimes-unveils-software-to-mimic-her-voice-and-announces-2-new-songs/
- TuneCore Help Center, "Distributing collaborations with GrimesAI": https://support.tunecore.com/hc/en-us/articles/16428915033492-Distributing-collaborations-with-GrimesAI
- Kits.ai, "Kits Earn": https://www.kits.ai/earn
- Universal Music Group / PR Newswire (Oct. 29, 2025): PR Newswire
- Warner Music Group (Nov. 19, 2025): WMG
- Warner Music Group (Nov. 25, 2025): WMG
- BBC News, "Warner Music Group settles lawsuit with AI music firm and launches joint venture" (Nov. 26, 2025): https://www.bbc.com/news/articles/cjdrl7lr039o
- Tennessee Office of the Governor, "Gov. Lee Signs ELVIS Act Into Law" (Mar. 21, 2024): https://www.tn.gov/governor/news/2024/3/21/photos--gov--lee-signs-elvis-act-into-law.html
- Tennessee General Assembly, HB 2091 (ELVIS Act) bill text: https://www.capitol.tn.gov/Bills/113/Bill/HB2091.pdf
- Office of the Governor of California, "Governor Newsom signs bills to protect digital likeness of performers" (Sept. 17, 2024): https://www.gov.ca.gov/2024/09/17/governor-newsom-signs-bills-to-protect-digital-likeness-of-performers/
- California Legislative Information, AB 2602 status: https://leginfo.legislature.ca.gov/faces/billStatusClient.xhtml?bill_id=202320240AB2602
- California Legislative Information, AB 1836 status: https://leginfo.legislature.ca.gov/faces/billStatusClient.xhtml?bill_id=202320240AB1836
- Congress.gov, S. 4591, NO FAKES Act of 2026: https://www.congress.gov/bill/119th-congress/senate-bill/4591
- Digital Music News, "Revised No Fakes Act Unanimously Advanced by the Senate Judiciary Committee" (June 18, 2026): https://www.digitalmusicnews.com/2026/06/18/no-fakes-act-senate-judiciary-advancement/
Part 3 coming soon.
