AI Voice Clones and the NO FAKES Act: What Creators Actually Own When Someone Copies Their Voice

Your voice is your brand. If you are a musician, a streamer, or any creator whose audience recognizes you by how you sound, AI voice cloning technology has crossed a threshold that should concern you. In 2023, a TikTok user calling themselves Ghostwriter977 uploaded "Heart on My Sleeve" — an AI-generated track mimicking the voices of Drake and the Weeknd — and it racked up millions of streams before Universal Music Group issued takedown notices. That was the wake-up call. Since then, AI voice cloning has only gotten better, cheaper, and more accessible. And until now, the legal tools available to creators whose voices have been cloned have been a patchwork of inconsistent state laws — many of which were never designed to address AI-generated replicas.

That may be about to change. The NO FAKES Act — formally the Nurture Originals, Foster Art, and Keep Entertainment Safe Act — was reintroduced in 2025 with bipartisan support and has been unanimously advanced by the Senate Judiciary Committee. If passed, it would create the first federal right of publicity specifically designed to address unauthorized AI-generated digital replicas of a person's voice or likeness. For creators, this is the most significant piece of federal legislation in a generation. Here is what it means, how it interacts with existing law, and what platforms are already doing about it.

The Current Patchwork: State Right-of-Publicity Laws

There is no federal right of publicity in the United States. Instead, the right to control the commercial use of your identity — including your voice — is governed by a patchwork of state statutes and common law that vary dramatically in scope, duration, and enforceability. Some states, like California and New York, have robust right-of-publicity statutes that cover voice imitation. Others have no statute at all, leaving creators to rely on common law torts like misappropriation or passing off.

The U.S. Copyright Office recognized this gap in its July 2024 report on digital replicas, recommending that Congress create a federal law to protect individuals from unauthorized uses of their digital replicas and noting the "inadequacy, narrowness, and inconsistency of existing laws." (Wikipedia, "NO FAKES Act," citing U.S. Copyright Office Report on Digital Replicas, July 2024)

This inconsistency creates a fundamental problem for creators whose audiences span every state. A voice clone that might be actionable in California could be perfectly legal in a state with no right-of-publicity statute. And because voice clones spread across the internet instantaneously, the question of which state's law applies — and whether any law applies at all — becomes a practical barrier to enforcement.

The Case Law Foundation: Midler and Waits

The legal precedent for protecting a distinctive voice against unauthorized imitation comes from two landmark Ninth Circuit cases from the late 1980s and early 1990s. Both were decided under state law, and both arose in the context of commercial advertising — not the broader universe of contexts where AI voice clones now appear.

In Midler v. Ford Motor Co., 849 F.2d 460 (9th Cir. 1988), Ford's advertising agency asked Bette Midler to sing a Ford commercial song. When she declined, they hired one of her backup singers to imitate her voice. The Ninth Circuit held that "a voice is as distinctive and personal as a face" and that the deliberate imitation of a distinctive voice for commercial purposes constituted a tort under California law. (Melendez v. Sirius XM Radio, Inc., 50 F.4th 294 (2d Cir. 2022), citing Midler v. Ford Motor Co., 849 F.2d 460 (9th Cir. 1988))

Four years later, in Waits v. Frito-Lay, Inc., 978 F.2d 1093 (9th Cir. 1992), the same court extended the principle when Frito-Lay used a Tom Waits soundalike in a Doritos radio commercial. Waits — who had also declined to license his voice — was awarded damages. The court reinforced that a distinctive voice is protectable when someone deliberately imitates it to capitalize on its recognizability. (Melendez v. Sirius XM Radio, Inc., 50 F.4th 294 (2d Cir. 2022), citing Waits v. Frito-Lay, Inc., 978 F.2d 1093 (9th Cir. 1992))

These cases established important principles, but they also reveal the gaps. Both involved commercial advertising. Both were decided under state law. And both required the plaintiff to prove that the defendant deliberately imitated a distinctive, well-known voice. An AI voice clone that goes viral on TikTok, appears in a YouTube video, or streams on Twitch does not fit neatly into the commercial advertising framework that Midler and Waits addressed. This is the gap the NO FAKES Act is designed to fill.

What the NO FAKES Act Would Do

The NO FAKES Act would establish a federal right of publicity that covers digital replicas — defined as "a newly created, computer-generated, highly realistic electronic representation that is readily identifiable as the voice or visual likeness of an individual," whether living or dead. (Wikipedia, "NO FAKES Act")

Here is what the bill would do if enacted:

Give every American — not just celebrities — rights in their voice and likeness. The Act grants right holders the exclusive right to authorize the use of their likeness in a digital replica. This is not limited to famous artists; it covers private individuals as well. The RIAA, which has strongly supported the bill, noted that it "give[s] every American a right in their voice & likeness from invasive AI deepfakes & voice clones — not just celebrities and other prominent figures." (RIAA Press Statement, June 2026)

Create a notice-and-takedown mechanism for platforms. Modeled on the DMCA's safe harbor framework, the Act would create a notice-and-takedown system under which online service providers must promptly respond to notifications seeking removal of unauthorized digital replicas. Platforms that adopt and enforce a policy for terminating accounts that repeatedly violate the Act would be shielded from liability — but the Act does not require proactive monitoring. (Wikipedia, "NO FAKES Act")

Protect First Amendment uses. The Act expressly protects uses for news reporting, parody, and satire — addressing concerns from digital rights organizations that a federal right of publicity could chill protected speech. (RIAA Press Statement, June 2026)

Establish postmortem rights. The right does not expire at death. It can be transferred to heirs or executors and lasts for 10 years following death, with potential extensions of 5 years — renewable up to 70 years post-mortem. (Wikipedia, "NO FAKES Act")

Hold both distributors and toolmakers liable. The Act establishes liability for individuals who knowingly distribute a digital replica without authorization, and for entities that make available a service primarily designed to produce unlawful digital replicas. (Wikipedia, "NO FAKES Act")

The Senate Judiciary Committee unanimously advanced the bill, with bipartisan support from Senators Blackburn (R-TN), Coons (D-DE), Tillis (R-NC), and Klobuchar (D-MN), among others. The White House's National Legislative Framework for AI has also endorsed the approach. (RIAA Press Statement, June 2026)

How Federal Legislation Interacts with State Law

One of the most important questions about the NO FAKES Act is how it would interact with existing state right-of-publicity laws. The Act is designed to create a federal floor — not to preempt state laws entirely. This means that states with stronger protections, like California, could continue to provide additional remedies beyond what the federal statute offers. But the federal cause of action would fill the gap in states that have no right-of-publicity statute or whose statutes do not address digital replicas.

For creators, the practical benefit is significant: instead of having to navigate 50 different state laws to determine whether your voice is protected, you would have a federal cause of action that applies nationwide. You could file in federal court. And the notice-and-takedown mechanism would give you a practical tool — sending a takedown notice — that does not exist under most state right-of-publicity frameworks today.

We have written about how creators can protect their work from AI training in our guide for visual artists on AI training opt-outs. The NO FAKES Act addresses a different but related concern: not who trained on your work, but who is generating content that sounds like you.

Why Musicians and Streamers Are Most Affected

Musicians and streamers face the most acute exposure to AI voice cloning. For musicians, the "Heart on My Sleeve" incident demonstrated that AI can now generate convincing vocals in a specific artist's style — and that the public will consume them eagerly. The track was pulled from streaming platforms after UMG issued takedown notices, but the legal mechanism used was reportedly the DMCA, not a right-of-publicity claim. (Wikipedia, "NO FAKES Act") The DMCA approach worked only because the track allegedly used an unauthorized sample — not because Drake or the Weeknd had a legal right to prevent AI imitation of their voices. Under current law, without a sample to point to, the takedown path is far less clear.

For streamers, voice is arguably the most important brand asset. Your audience tunes in for your commentary, your reactions, your personality — all of which are expressed through your voice. An AI clone that generates offensive content in your voice, or that impersonates you on another platform, can damage your reputation and your sponsor relationships. We have written about the contractual side of brand deals in our guide to brand deal contracts for creators, and about AI-related liability in our analysis of AI product liability and insurance gaps. Voice cloning sits at the intersection of both: it threatens both your IP and your brand partnerships.

What Platforms Currently Require

While the NO FAKES Act would create a federal legal framework, the platforms where voice clones actually spread have already begun implementing their own policies. Here is what YouTube, TikTok, and Twitch currently require.

YouTube

YouTube requires creators to disclose AI-generated or altered content that realistically depicts a real person doing or saying something they did not do. The platform introduced its AI-generated content disclosure policy in March 2024, requiring creators to use a label indicating that content was altered or generated by AI. YouTube can remove AI-generated content that impersonates a real person, and it provides a privacy request process for individuals whose likeness has been used without consent. (YouTube Blog, "An update on our approach to AI-generated content")

TikTok

TikTok requires creators to label AI-generated content that contains realistic images, audio, or video. The platform prohibits AI-generated content that shows the likeness of young people under the age of 18, or the likeness of adult private figures used without their permission. TikTok also offers both creator-applied labels and automatic labeling through C2PA Content Credentials. Content that depicts AI-generated speech — including voice clones — must be labeled under TikTok's Community Guidelines. (TikTok Help Center, "About AI-generated content")

Twitch

Twitch addresses deepfake and synthetic content through its Community Guidelines, which prohibit content that depicts someone without their consent — including deepfakes and synthetic media. Twitch has a dedicated policy addressing explicit deepfake content, and its broader impersonation rules extend to AI-generated content that mimics a real person. (Twitch Community Guidelines)

These platform policies are important, but they are not law. They are terms of service — contractual obligations between the platform and its users. A platform can remove a voice clone, suspend an account, or demonetize content, but it cannot award damages. The NO FAKES Act would provide the legal cause of action that platform policies cannot: the right to sue the person who created or distributed the unauthorized voice clone.

Actionable Next Steps

1. Document your voice as a brand asset. If your voice is central to your career — as a musician, streamer, or podcaster — treat it as intellectual property. Record samples of your work, document where and how your voice is used commercially, and keep records of any sponsorships or licensing deals tied to your vocal identity. This documentation is the foundation for any future enforcement action.
2. Monitor for unauthorized voice clones. Set up alerts for your name and common variations on YouTube, TikTok, and Twitch. If you find content that uses an AI-generated voice resembling yours, document it immediately — screenshot the post, record the URL, and note the date. This evidence is essential for both platform takedowns and any future legal action under the NO FAKES Act or state law.
3. Use platform takedown tools now. You do not need to wait for the NO FAKES Act to act. YouTube's privacy request process, TikTok's reporting system, and Twitch's Community Guidelines enforcement all provide mechanisms for requesting removal of unauthorized AI-generated content that impersonates you. Use them — and document your requests.
4. Register your copyrights. If your voice appears in original works — songs, recordings, videos — register those works with the U.S. Copyright Office. As we explain in our guide for creators on AI and IP protection, registration within three months of publication preserves your right to statutory damages and attorneys' fees. While copyright does not directly protect your voice as an identity asset, it protects the works in which your voice appears — and that is often the most practical enforcement lever available today.
5. Follow the NO FAKES Act's progress. The bill has been unanimously advanced by the Senate Judiciary Committee and has bipartisan support in both chambers. If it passes, it will create a federal cause of action and a notice-and-takedown mechanism that does not currently exist. Stay informed about its progress so you can act quickly when it becomes law.
6. Review your contracts for voice licensing terms. If you have brand deals, sponsorships, or distribution agreements, check whether they address AI voice cloning. Do they prohibit the brand from creating AI voice clones of you? Do they address who owns the right to authorize digital replicas? As we discuss in our brand deal contract guide for creators, the contracts you sign today determine what rights you retain — and what you give away.
7. Engage counsel before you need enforcement. If you discover an unauthorized voice clone, you need to act quickly — but you also need to act correctly. The legal landscape is shifting rapidly, and the intersection of state right-of-publicity law, platform policies, and the forthcoming NO FAKES Act creates a complex enforcement picture. An attorney who understands creator IP, AI law, and the specific platforms where you operate can help you build a strategy before the clone goes viral.

The NO FAKES Act represents the most significant expansion of creator rights in the AI era. It would give every American — not just celebrities — a federal right to control unauthorized AI-generated replicas of their voice. Until it passes, creators must navigate a patchwork of state laws and platform policies that are incomplete, inconsistent, and often inadequate. The creators who document their voice assets, monitor for clones, use existing takedown tools, and build a legal strategy now will be the ones who can act quickly when the law catches up to the technology. The ones who wait will find that their voice has already been copied — and that the legal tools to get it back are still being written.