AI Voice Cloning and Music Copyright: A Dual-Risk Guide for Indie Musicians in 2026

AI voice cloning has moved from novelty to existential threat for working musicians. In June 2024, the RIAA sued Suno and Udio — the two largest AI music generation platforms — alleging they trained their models on copyrighted sound recordings without licenses. Tennessee passed the ELVIS Act, the first state law explicitly protecting musicians' voices from AI cloning. The U.S. Copyright Office released its AI reports addressing digital replicas and copyrightability. And the D.C. Circuit confirmed that AI-generated works without human authorship cannot be copyrighted.

If you're an indie musician, you face a dual risk: someone cloning your voice without permission, and you losing rights when you use AI tools in your own production. This guide breaks down both sides and gives you actionable contract clauses to protect yourself in 2026.

The Two-Sided AI Threat to Indie Musicians

AI voice cloning and AI music copyright risk operate on two distinct fronts. On one side, bad actors can use AI to replicate your voice, your likeness, and your sound — creating unauthorized tracks that dilute your brand and siphon your audience. On the other side, when you use AI platforms like Suno or Udio to generate music, you may be signing away rights you don't realize you're granting — and the output itself may not be copyrightable at all.

We've covered the basics of AI voice cloning and musician rights before, and explored what musicians actually own when they use AI music tools. This article brings both sides together into a single, practical framework — including the producer agreement clauses you need to negotiate now.

How Your Voice Is (and Isn't) Protected Today

Your voice sits at the intersection of two legal frameworks: state right-of-publicity laws and federal copyright for sound recordings. Neither is complete on its own, and the gaps between them are where AI cloning threats live.

State Right-of-Publicity Laws: The ELVIS Act and Beyond

Right-of-publicity laws protect your ability to control the commercial use of your identity — your name, image, likeness, and in some states, your voice. But until recently, most state statutes were written long before AI voice cloning existed.

Tennessee's ELVIS Act (Ensuring Likeness Voice and Image Security Act), codified at T.C.A. § 47-25-1101 et seq., changed that. Signed into law on March 21, 2024, and effective July 1, 2024, the ELVIS Act amended Tennessee's existing right-of-publicity statute to explicitly protect a person's voice from AI-generated cloning. As described in a detailed analysis by Proskauer, the Act creates a novel cause of action against anyone who distributes a technology whose "primary purpose or function" is creating a facsimile of a person's voice without authorization — potentially subjecting AI tool providers themselves to liability, not just the end users who generate cloned audio.

New York has also moved aggressively. In December 2025, Governor Hochul signed two landmark laws: one requiring disclosure of AI-generated "synthetic performers" in commercial advertisements (effective June 2026), and another strengthening posthumous right-of-publicity protections by requiring prior consent from heirs before commercially using a deceased individual's name, voice, image, or likeness — including AI-generated "digital replicas" of deceased performers. According to Debevoise & Plimpton's analysis, violations can result in statutory damages of $2,000 or compensatory damages (whichever is greater), plus punitive damages.

Illinois has long had one of the strongest right-of-publicity statutes — the Right of Publicity Act, 765 ILCS 1075 — which protects an individual's "name, voice, signature, photograph, likeness, or identity" for 50 years after death. Illinois has also considered legislation (HB 4875) to add explicit digital replica protections, signaling that the state is paying attention to AI-specific voice threats.

The problem? Right-of-publicity laws are state-by-state. There is no federal right of publicity. If you live in a state without a strong statute — or if the person cloning your voice operates from outside the U.S. — your protections may be thin. That's where federal copyright comes in.

Federal Copyright for Sound Recordings

Federal copyright law protects sound recordings — the specific recorded performance of a song — separately from the underlying musical composition. This is a distinction indie musicians often miss. The composition (the melody, lyrics, and chord progression) is one copyright, typically held by the songwriter or publisher. The sound recording (the actual audio file, the mastered track) is a separate copyright, typically held by the artist or label.

When someone clones your voice, they may not be copying a specific recording — they're generating new audio that sounds like you. This means traditional copyright infringement claims may not cleanly apply, because copyright protects expression, not a person's vocal characteristics. The U.S. Copyright Office's AI reports — Part 1 on digital replicas (published July 31, 2024) and Part 2 on copyrightability (published January 29, 2025) — address this gap. Part 1 specifically recommends a federal digital replica law to protect against unauthorized AI-generated likenesses, acknowledging that existing copyright law alone is insufficient.

For your existing recordings, copyright registration remains essential. If someone copies or distributes your actual recorded tracks to train an AI model, that's a direct copyright infringement claim — which is exactly what the RIAA alleged against Suno and Udio.

The RIAA v. Suno and Udio Litigation: Why It Matters for You

On June 24, 2024, the RIAA — on behalf of Sony Music, Universal Music Group, and Warner Music Group — filed separate lawsuits against Suno (in the District of Massachusetts) and Udio (in the Southern District of New York). The complaints allege that both platforms trained their AI models on copyrighted sound recordings without obtaining licenses, seeking statutory damages of up to $150,000 per infringed work.

According to reporting on the case, Suno admitted in its August 2024 court filings that it trained on copyrighted recordings but argued the use was transformative fair use. Warner Music Group settled with Suno in November 2025 and signed a licensing deal, but Sony Music and Universal Music Group continue to litigate — and have moved to expand the case to cover over 61,000 recordings. This litigation will shape the AI music industry for years, and its outcome directly affects what platforms can offer you and what risks you face when you use them.

For indie musicians, the RIAA litigation signals two things. First, the industry is treating unauthorized AI training on copyrighted recordings as a serious legal threat. Second, the eventual outcomes — whether through court rulings or settlements — will define what "licensed" AI music platforms look like going forward. As we've discussed in our guide to AI voice clones and the NO FAKES Act, the legal landscape is shifting rapidly.

What Happens When You Use AI Music Tools (Suno, Udio)

The flip side of protecting your voice from cloning is understanding what you risk when you use AI tools in your own production. Two issues dominate: platform terms of service and human authorship requirements under copyright law.

Platform Terms of Service: Who Owns What?

When you generate music on Suno or Udio, the platform's terms of service govern what you can do with the output. Suno's current Terms of Service (last revised March 26, 2026) state that songs downloaded from Suno on paid plans "remain yours to use commercially or personally." However, free-tier users only get "trial downloads" that are for personal use — commercial rights require a paid subscription.

Suno's August 2026 blog post announcing upcoming terms changes confirmed this structure: "For any song that you download from the platform as a paying subscriber, you have the commercial rights to the music and can use it for commercial or personal use." The company also introduced download limits (Free: 7 lifetime; Pro: 20/month; Premier: 60/month) and announced new models "developed in partnership with the music industry."

But here's the catch: "commercial rights" under a platform's terms of service is not the same as owning a copyright. Suno granting you commercial use rights doesn't mean the U.S. Copyright Office will register the work. And if Suno's own training data was infringing — as the RIAA lawsuit alleges — your "commercial rights" could be undermined if the platform is forced to change its model or pull content.

Suno's terms also include a notable Voice Model provision: you can only create a voice model of your own voice. The terms explicitly state that users agree "not to create, or attempt to create, a voice model of another person." This is a platform-level guardrail, but it's not a substitute for legal protection — and it only binds Suno users, not the broader ecosystem of AI cloning tools.

Udio's terms of service are less accessible, but similar concerns apply. The broader principle: read the ToS carefully before you upload anything, and understand that platform-granted rights and federal copyright are two different things.

Human Authorship and Copyrightability After Thaler

Even if a platform grants you commercial rights to AI-generated music, you may not be able to copyright it. The D.C. Circuit Court of Appeals confirmed in Thaler v. Perlmutter (March 2025) that human authorship is required for copyright protection under U.S. law. The Supreme Court subsequently denied certiorari, making this the final word — for now.

The Copyright Office's Part 2 report, published January 29, 2025, elaborated on this principle for generative AI outputs. The Office concluded that works generated entirely by AI — without sufficient human creative contribution — are not copyrightable. However, works where a human exercises sufficient creative control (such as selecting, arranging, or modifying AI-generated elements) may be registrable, with the human-authored portions protected.

For indie musicians, this means: if you type a text prompt into Suno and download the result, you likely cannot copyright the output. If you take that output, re-record vocals over it, rearrange the structure, and mix it with your own original instrumentation, the human-authored elements may be protectable — but the AI-generated portions may not be. This matters enormously for licensing, sampling clearance, and protecting your work from being copied by others.

Producer Agreement Clauses You Need in 2026

If you work with producers, beatmakers, or collaborators who use AI tools, your contracts need to address AI-specific risks. Standard producer agreements from even two years ago are likely insufficient. Here are the clauses you should negotiate:

1. AI Disclosure and Authorization Clause

Require your producer to disclose any use of AI tools (Suno, Udio, AI voice models, AI stem separators) in the creation of any deliverable. The clause should specify that unauthorized AI-generated content constitutes a material breach. This protects you from unknowingly releasing music with uncopyrightable elements or content that may infringe third-party rights.

2. Voice Clone Prohibition

Explicitly prohibit the producer from using AI to clone, replicate, or simulate your voice — or any third party's voice — without written consent. Reference applicable state laws (the ELVIS Act if you have Tennessee connections, New York's digital replica statutes, Illinois's Right of Publicity Act). This is especially critical if you share vocal stems or acapellas with producers during collaboration.

3. AI Training Data Warranty

Require the producer to warrant that any AI tools used were trained on properly licensed data, and to indemnify you against claims arising from AI-generated content that infringes third-party copyrights. Given the active RIAA v. Suno litigation, this indemnity is non-negotiable — if a platform's training data is found to be infringing, you need contractual recourse.

4. Copyright Registration Cooperation Clause

Specify that any copyright registration for collaborative works will identify which elements are human-authored and which are AI-generated, consistent with Copyright Office guidance. The clause should require the producer to cooperate in preparing registration applications that accurately reflect authorship — because misrepresenting AI-generated content as human-authored can result in cancellation of your registration.

5. Sound Recording vs. Composition Split

Ensure your agreement clearly distinguishes between the sound recording copyright and the underlying musical composition copyright — and specifies how AI-generated elements affect each. If a producer used Suno to generate a beat, the composition may be partially AI-generated and potentially uncopyrightable, even if your vocals and lyrics on top are fully human-authored and protectable.

6. Reversion and Termination Rights

Include a clause that allows you to terminate the agreement and reclaim rights if AI-generated content in the deliverables is later found to infringe third-party rights or if the underlying AI platform is shut down, compelled to change its terms, or loses litigation that affects the validity of outputs.

Actionable Next Steps

Here's what you can do this month to protect your voice, your recordings, and your brand:

1. Register your sound recordings with the U.S. Copyright Office. Registration is a prerequisite for filing an infringement lawsuit and provides statutory damages and attorney's fees. If you haven't registered your published recordings, do it now.
2. Audit your producer and collaboration agreements. If they don't include the AI-specific clauses above, request an addendum. Don't wait until a dispute arises to discover that your contract doesn't address AI-generated content.
3. Document your creative process. Keep records of your original recordings, vocal takes, and production decisions. If you ever need to prove human authorship for copyright registration — or prove that someone cloned your voice — a clear creative record is your best evidence.
4. Read AI platform terms of service before you upload anything. Understand what rights you're granting, what rights you're receiving, and whether "commercial use" means what you think it means. If you're using a free tier, you almost certainly don't have commercial rights.
5. Monitor for unauthorized voice clones. Set up alerts for your name and artist handle across streaming platforms and social media. If you discover unauthorized AI-generated content using your voice, document it immediately and consult counsel about takedown options under the DMCA, right-of-publicity laws, and platform reporting mechanisms.
6. Consult a music attorney who understands AI law. The intersection of right-of-publicity statutes, federal copyright, and AI platform terms is complex and evolving. Generic legal templates won't protect you — you need counsel who tracks the RIAA litigation, state legislative developments, and Copyright Office guidance.