AI Voice Cloning and Musician Rights: Right of Publicity, the NO FAKES Act, and Platform Takedowns in 2026

In April 2023, a TikTok user known as Ghostwriter977 uploaded "Heart on My Sleeve," a song that sounded remarkably like a collaboration between Drake and The Weeknd. It racked up millions of streams before Universal Music Group issued takedown notices to Spotify, YouTube, TikTok, and Apple Music. The catch? Neither Drake nor The Weeknd had anything to do with the track. Their voices were cloned using AI. The "Heart on My Sleeve" incident was a wake-up call for the music industry — and it was only the beginning.

Since then, AI voice cloning tools from Suno, Udio, ElevenLabs, and others have made it possible for anyone with a laptop to generate convincing imitations of any musician's voice. The legal landscape is racing to catch up. In June 2024, the RIAA filed landmark copyright infringement lawsuits against Suno and Udio. Tennessee updated its right-of-publicity law to specifically address AI voice clones. California, Illinois, and New York have enacted or are advancing their own digital replica legislation. And at the federal level, the NO FAKES Act — which would create the first federal right of publicity — has been reintroduced with bipartisan support.

If you're a musician, your voice is your most valuable asset. Here's what the law says about protecting it in 2026 — and what you can actually do when someone clones it without permission.

For related reading on how AI is reshaping creative rights, see our guide on what game studios actually own when AI generates their assets and our analysis of AI product liability and insurance gaps in 2026.

The Federal Gap: Why There's No National Right of Publicity

The United States has no federal right of publicity. Instead, the right to control the commercial use of your name, image, likeness, and voice is governed by a patchwork of state statutes and common law — and the protections vary dramatically depending on where you live and where the infringement occurs. Some states, like California, have robust statutory rights of publicity. Others recognize only narrow common law protections. Some extend rights posthumously for decades; others expire them at death.

This fragmentation creates a fundamental problem for musicians: an AI voice clone uploaded from one state, hosted on servers in another, and streamed globally may fall through the cracks of multiple jurisdictions' laws. As the U.S. Copyright Office noted in its July 2024 report on digital replicas, existing state laws are "inadequate, narrow, and inconsistent" in addressing the threat posed by AI-generated voice and likeness replicas. The Office recommended that Congress enact a federal law to protect individuals from unauthorized digital replicas.

The NO FAKES Act: A Federal Right of Publicity on the Horizon

The Nurture Originals, Foster Art, and Keep Entertainment Safe Act — the NO FAKES Act — is the most significant federal legislative response to AI voice cloning. First introduced as a discussion draft in 2023, formally introduced in 2024, and reintroduced in April 2025 by Senators Coons, Blackburn, Klobuchar, and others, the bill would establish the first federal right of publicity specifically targeting digital replicas.

Under the NO FAKES Act, a "digital replica" is defined as "a newly created, computer-generated, highly realistic electronic representation that is readily identifiable as the voice or visual likeness of an individual," living or dead. The Act would grant right holders the exclusive authority to authorize the use of their likeness in a digital replica, creating a licensing framework similar to existing copyright licensing. The bill also establishes liability for individuals who knowingly distribute unauthorized digital replicas and for entities that operate services primarily designed to produce them.

Crucially, the NO FAKES Act includes a notice-and-takedown mechanism modeled on the DMCA — something current state right-of-publicity laws lack. Online service providers would be shielded from liability if they adopt and enforce a policy for terminating accounts that repeatedly violate the Act and designate an agent with the U.S. Copyright Office to receive takedown notifications. This is a game-changer for musicians: it would give you a federal tool to demand removal of AI voice clones from platforms, without needing to navigate the limitations of copyright law or the variability of state statutes.

The Act also provides postmortem rights lasting 10 years after death, with possible 5-year extensions up to 70 years — a significant protection for the estates of deceased artists whose voices could be cloned from existing recordings.

The NO FAKES Act has drawn broad support from the entertainment industry, including the RIAA, SAG-AFTRA, and the Recording Academy. However, digital rights organizations have raised concerns about potential chilling effects on protected speech, and the bill remains in committee as of 2026. The National Law Review reports that the reintroduced version has attracted broader industry support, but passage is not guaranteed.

State AI Likeness Laws: A Growing Patchwork

While Congress debates the NO FAKES Act, states are not waiting. Several have already enacted legislation specifically targeting AI-generated voice and likeness replicas, and more are considering bills in 2025-2026.

Tennessee's ELVIS Act: The First State to Target AI Voice Cloning

Tennessee became the first state to enact AI-specific right-of-publicity protections when Governor Bill Lee signed the Ensuring Likeness, Voice, and Image Security (ELVIS) Act into law on March 21, 2024. The legislation, House Bill 2091, amended Tennessee's existing right of publicity statute to explicitly cover "a voice" as a protected element of a person's likeness and to address AI-generated reproductions.

The updated ELVIS Act prohibits the use of an individual's voice or likeness in a "digital replica" without consent. It defines a digital replica as a computer-generated, highly realistic electronic representation that is readily identifiable as the individual's voice or visual likeness. The law applies to both living and deceased individuals — Tennessee's posthumous right of publicity extends for 10 years after death, consistent with the state's long history of protecting the legacy of musicians like Elvis Presley.

For musicians based in or performing in Tennessee — which includes much of Nashville's music industry — the ELVIS Act provides a direct state-law cause of action against unauthorized AI voice cloning. The law allows for injunctive relief, damages, and attorney's fees.

California AB 2602: Consent Requirements for Digital Replicas in Contracts

California, home to much of the recording industry, enacted Assembly Bill 2602 in September 2024. AB 2602 renders contractual provisions unenforceable if they allow the creation and use of a digital replica of an individual's voice or likeness in place of work the individual would otherwise have performed in person — unless three conditions are met: (1) the contract provides a "reasonably specific" list of all proposed uses of the digital replica, (2) the individual is represented by legal counsel or a labor union during negotiation, and (3) the uses are specifically described.

This law directly addresses a common industry abuse: record labels and studios inserting broad language like "all media known or hereafter devised" into contracts to claim rights over AI-generated replicas. Under AB 2602, that boilerplate language is no longer sufficient to secure digital replica rights. The law took effect for performances fixed on or after January 1, 2025.

California also enacted AB 1836, which prohibits the creation or distribution of digital replicas of deceased personalities without consent from their estate. Noncompliance carries penalties of $10,000 or actual damages, whichever is greater.

Illinois: Digital Voice and Likeness Protection

Illinois amended its right of publicity law through House Bill 4762, which was signed into law in 2024. The amendment expanded the state's existing right of publicity protections to explicitly cover digital replicas of voices and likenesses, including those created using AI. Illinois already had one of the strongest right-of-publicity statutes in the country, and the amendment modernized it to address the specific threat of AI voice cloning.

New York: Landmark AI Right of Publicity Law

New York enacted its own landmark AI right of publicity legislation, signed in late 2025. According to Debevoise & Plimpton's analysis, the new law creates protections for living and deceased individuals against unauthorized digital replicas, with specific provisions covering AI-generated voice and likeness reproductions. The law requires disclosure when AI-generated performers are used in advertising and creates penalties for unauthorized commercial use of digital replicas.

The practical implication for musicians: if your voice is cloned and the resulting track is commercially exploited in California, Tennessee, Illinois, or New York, you now have a state-law cause of action — even without a federal right of publicity. But the specific remedies, definitions, and procedural requirements vary by state, and a track streamed nationally may require analysis under multiple statutes.

Recent Litigation: RIAA v. Suno and RIAA v. Udio

On June 24, 2024, the Recording Industry Association of America announced the filing of two landmark copyright infringement lawsuits against AI music generation services. The case against Suno, Inc. was filed in the U.S. District Court for the District of Massachusetts, and the case against Uncharted Labs, Inc. (developer of Udio) was filed in the Southern District of New York. Plaintiffs include Sony Music Entertainment, UMG Recordings, and Warner Records.

The complaints allege that Suno and Udio copied decades of copyrighted sound recordings — without permission — to train their AI models. According to the RIAA's press release, the complaints describe "copying decades worth of the world's most popular sound recordings and then ingesting those copies to generate outputs that imitate the qualities of genuine human sound recordings." The lawsuits seek declarations of infringement, injunctions barring future unauthorized use, and damages.

These cases are significant for musicians for two reasons. First, they target the training-data side of AI voice cloning — the mass ingestion of recordings that enables AI models to learn and reproduce specific artists' vocal characteristics. Second, they test whether the fair use doctrine shields AI companies that ingest copyrighted works for model training. Suno and Udio have argued that their training practices constitute fair use, but the RIAA's complaints directly contest this, arguing that the services "offer imitative machine-generated music — not human creativity or expression."

The outcomes of these cases will shape how AI music tools can be built and trained — and whether record labels and artists can control how their recordings are used to train models that might eventually clone their voices.

The "Heart on My Sleeve" Precedent and Platform Takedowns

When UMG issued takedown notices for "Heart on My Sleeve," the legal basis was not entirely clear. According to reporting on the incident, UMG likely relied on the DMCA, based on an unauthorized sample used at the start of the song for a producer tag. But the core issue — the AI imitation of Drake's and The Weeknd's voices — did not fit neatly into existing copyright law, because the song was an original composition that merely sounded like the artists. UMG did not hold copyright in the artists' voices themselves.

This gap between copyright law (which protects specific recordings and compositions) and right-of-publicity law (which protects an individual's identity) is exactly what the NO FAKES Act and state AI likeness laws are designed to fill. Until those laws provide effective notice-and-takedown mechanisms, musicians are left to navigate a system that was not designed for AI voice cloning.

Grimes's Open-Source Voice Licensing Experiment

Not every artist sees AI voice cloning as a threat. In 2023, Grimes launched Elf.Tech, a platform that allowed creators to generate music using an AI model of her voice — and offered a 50/50 royalty split on any successful tracks. As the BBC reported, Grimes invited creators to use her voice "without penalty" as long as they registered their tracks through her platform for royalty distribution.

Grimes's experiment represents the other side of the AI voice cloning debate: the possibility that musicians could proactively license their voice for AI use, maintaining control and generating revenue. But it also highlights the legal complexity. Grimes was able to authorize this use because she owns her own master recordings and has the rights to her voice. Many artists under label contracts may not have that same autonomy — and as California's AB 2602 makes clear, any licensing of digital replica rights must be specific, informed, and represented by counsel.

Platform Takedown Strategies: DMCA vs. Right of Publicity

When you discover that someone has cloned your voice and uploaded the result to YouTube, TikTok, Spotify, or SoundCloud, your options for removal depend on the platform's policies and the legal theory you invoke. Here's what works — and what doesn't — in 2026.

The DMCA Approach: Limited but Familiar

The Digital Millennium Copyright Act provides a notice-and-takedown mechanism for copyright infringement. If an AI-generated track incorporates an actual sample of your copyrighted recording — even a brief one — you can file a DMCA takedown notice with the platform. This is likely what UMG did with "Heart on My Sleeve," which reportedly included an unauthorized sample at the beginning.

The DMCA's advantage is speed: platforms are generally required to act expeditiously to remove infringing content after receiving a valid takedown notice. The disadvantage is that the DMCA only covers copyright infringement — not right of publicity violations. If an AI track merely mimics your vocal style without incorporating an actual sample of your recording, the DMCA may not apply. You can't file a DMCA takedown for "this song sounds like me" — you need to identify specific copyrighted material that was reproduced.

Right of Publicity Claims: Stronger but Slower

If your voice has been cloned without an actual sample being used, your strongest legal basis is a right of publicity claim under state law. Tennessee's ELVIS Act, California's AB 2602, Illinois's amended statute, and New York's new law all provide causes of action for unauthorized digital voice replicas. The problem is that state right of publicity laws do not include the same streamlined notice-and-takedown mechanisms as the DMCA. You typically need to send a cease-and-desist letter or file a lawsuit — a slower and more expensive process.

However, platforms are increasingly responsive to right of publicity complaints, even without a formal statutory takedown mechanism. YouTube's policies prohibit impersonation, and the platform has taken down AI-generated content that mimics real artists. TikTok has similarly updated its policies to address AI-generated content. Spotify and SoundCloud have policies against unauthorized use of artists' names and likenesses. The key is framing your complaint correctly: cite the specific state statute that applies, document that the content is an AI-generated digital replica of your voice, and request removal under the platform's terms of service in addition to any statutory claim.

The NO FAKES Act's Promise: A Federal Takedown Mechanism

If enacted, the NO FAKES Act would create the most powerful takedown tool for musicians: a federal notice-and-takedown mechanism specifically for digital replicas. You would be able to send a notification to a platform's designated agent, and the platform would be required to promptly remove the unauthorized digital replica to maintain its safe harbor protection. This would eliminate the gap between the DMCA's copyright-only scope and the slow, state-by-state right of publicity process.

Licensing Your Voice for AI: What Agreements Must Include

If you choose to license your voice for AI use — whether through a platform like Grimes's Elf.Tech, a direct deal with an AI company, or a provision in a record label contract — the agreement must be drafted carefully. California's AB 2602 provides a useful framework for what a compliant digital replica license should include, even if you're not based in California.

• Specific description of uses: The agreement must list every proposed use of the digital replica. "All media known or hereafter devised" is not sufficient. Specify whether the replica will be used for new recordings, live performances, advertising, film, or training data.
• Scope and duration: Define how long the license lasts and whether it's exclusive or non-exclusive. The NO FAKES Act proposes that licenses for living individuals be limited to 10 years, with the right being non-assignable during the individual's lifetime.
• Compensation structure: Specify royalty rates, advance payments, and how revenue from AI-generated content will be calculated and shared. Grimes's 50/50 model is one approach, but the terms should be negotiated, not defaulted to.
• Quality control: Include provisions that give you approval rights over the quality and nature of the AI-generated output. You should have the right to reject uses that could damage your reputation.
• Revocation rights: Specify the conditions under which you can revoke the license — for example, if the AI company uses your voice for purposes beyond the agreed scope, or if the technology produces outputs that misrepresent you.
• Postmortem rights: Address what happens to the license if you die. State laws vary on posthumous right of publicity duration, and the NO FAKES Act would establish a federal framework for postmortem digital replica rights.
• Legal representation: Under California's AB 2602, if you're licensing digital replica rights in a personal or professional services contract, you must be represented by counsel or a labor union during negotiation. Even if you're not in California, having an attorney review any AI voice licensing agreement is essential — these are novel contracts with significant IP and reputational implications.

Actionable Next Steps

1. Register your voice as a trademark if it functions as a brand identifier. While copyright protects specific recordings, trademark law can protect your voice if it has become distinctive enough to identify you as the source of goods or services. Consult with an IP attorney about whether your vocal signature qualifies for trademark protection — this adds a federal enforcement layer on top of state right of publicity laws.
2. Review your existing recording contracts for digital replica language. Check whether your label or publisher has included broad language that could be construed as granting rights to create AI digital replicas of your voice. Under California's AB 2602, such provisions are unenforceable unless they meet specific consent and representation requirements — but you should know what your contracts say before a dispute arises.
3. Monitor platforms for unauthorized voice clones. Set up alerts on YouTube, TikTok, Spotify, and SoundCloud for your name and variations of it. When you find an unauthorized AI track, document it — screenshot the upload, record the URL, note the platform's takedown procedures, and preserve evidence of commercial exploitation (streams, ads, monetization).
4. File takedowns strategically. If the unauthorized track contains an actual sample of your recording, file a DMCA takedown notice — it's the fastest path. If it's a pure AI imitation without a sample, invoke the platform's impersonation and right of publicity policies, cite the applicable state statute (Tennessee's ELVIS Act, California's AB 2602, etc.), and send a cease-and-desist letter to the uploader. Consult with an attorney before filing to make sure your notice is legally sound.
5. Advocate for the NO FAKES Act. The federal notice-and-takedown mechanism proposed in the NO FAKES Act would be the single most powerful tool for musicians fighting AI voice cloning. Contact your representatives and urge them to support the bill — industry organizations like the RIAA and the Recording Academy are already lobbying for it, but individual artist voices matter.
6. If you want to license your voice for AI, get it in writing. Whether you're working with an AI company, a platform, or your own label, any agreement that touches your digital replica rights must be specific about scope, duration, compensation, quality control, and revocation. Don't sign a contract that grants AI voice rights without legal review.
7. Get legal counsel who understands AI and music law. The intersection of right of publicity, copyright, and AI is evolving rapidly. Generic legal advice won't protect you. You need an attorney who understands the specific state statutes, the pending federal legislation, and the platform takedown landscape. Book a consultation with our team — we work with musicians and creators on AI rights, IP protection, and platform enforcement every day.

AI voice cloning is not going away. The tools will get better, cheaper, and more accessible. The legal landscape will continue to evolve — but it's moving in the direction of stronger protections for musicians. The question is whether you'll be prepared when someone clones your voice. Understanding your rights under state and federal law, knowing how to use platform takedown mechanisms, and having contracts that properly license your digital replica rights are the three pillars of protecting your most valuable asset in the AI era.