The NO FAKES Act — formally the Nurture Originals, Foster Art, and Keep Entertainment Safe Act — is proposed United States federal legislation designed to curb unauthorized digital replicas of a person's voice and likeness. For voice actors, it represents the most serious attempt yet to create a federal publicity right that treats an AI-cloned voice as property the performer controls, rather than something anyone with a voice model and a few minutes of audio can freely exploit. As of August 2026, the bill has cleared the Senate Judiciary Committee but has not become law, which means voice actors today still operate in a patchwork of state right-of-publicity laws, union agreements, and contract terms.
What the NO FAKES Act Actually Does
Also worth reading: How can creators protect their digital vocal identity from AI cloning and misuse in 2026? · What are the legal and ethical standards for AI voice cloning consent for podcasters in 2026? · How does AI voice cloning work for podcast editing in 2026, and is it worth the risk?
The core mechanism of the NO FAKES Act is simple: it would give individuals a federal property right in their own voice and likeness, allowing them to authorize or refuse digital replicas. A "digital replica" under the bill is defined broadly enough to cover AI-generated voice performances that imitate a specific person's distinctive vocal characteristics — not just literal audio samples. If enacted, unauthorized replication could trigger statutory damages, actual damages, profits attributable to the infringement, and attorney's fees, giving performers a real financial deterrent against bad actors.
The bill also addresses post-mortem rights, extending protection for a period after death (with provisions for rights holders to maintain or terminate those protections), which matters enormously for estates of iconic voice actors whose voices are frequently cloned without permission. Importantly, the legislation includes exemptions intended to balance free expression: First Amendment protections for documentaries, parody, satire, criticism, and news reporting remain intact. This means the law targets commercial impersonation — fake endorsements, cloned narration sold as if performed by the real actor — rather than legitimate commentary.
For working voice actors, the practical effect would be a single federal standard replacing the current mess of state-by-state rules. Right now, whether you have recourse depends heavily on where you live and where the infringer operates; states like California and Tennessee (which passed its ELVIS Act specifically covering voice in 2024) offer strong protection, while many states offer little or none.
Why Voice Actors Are Pushing So Hard
The urgency behind the NO FAKES Act comes from how cheap and convincing voice cloning has become. Modern tools can produce a usable clone from seconds to a few minutes of reference audio, and synthetic speech quality has crossed the threshold where casual listeners cannot reliably distinguish it from human performance. For voice actors, whose entire professional identity is their voice, this is an existential threat: studios, game developers, and advertisers can theoretically license a clone once and never hire the performer again.
SAG-AFTRA has made this a headline issue. More than 16,000 actors signed a SAG-AFTRA letter urging Congress to pass the NO FAKES Act, one of the largest coordinated advocacy efforts in the union's recent history. The union's position is that consent and compensation must precede any use of a performer's voice — the same principle that drove the 2023 strike settlements, which included provisions on informed consent and compensation for digital replicas created during employment.
High-profile incidents have kept pressure on lawmakers. Taylor Swift's response to AI-generated voice deepfakes — including filing trademark applications to protect her voice and likeness — demonstrated that even the most powerful celebrities lack complete legal tools today. If someone at that level needs new legal instruments, working voice actors with far fewer resources need them even more.
The Current Legal Landscape Without the Act
Until the NO FAKES Act passes both chambers and is signed into law, voice actors must rely on existing protections, each with gaps. State right-of-publicity laws vary widely: some states recognize publicity rights only for famous individuals, some only through common law, and several recognize none at all. Tennessee's ELVIS Act, signed in March 2024, was the first state statute explicitly naming "voice" as protected, making unauthorized AI voice cloning actionable there regardless of fame.
Copyright law offers surprisingly little help. Courts have generally held that a voice itself is not copyrightable subject matter — copyright protects fixed works, not the timbre of a person speaking. A voice actor may hold copyright in a specific recorded performance, but not in the sound of their voice generally, which is why cloning a voice from publicly available recordings often falls outside copyright entirely.
Contract law fills some gaps. Union agreements negotiated by SAG-AFTRA now include consent and compensation requirements for digital replicas made from covered work. Individual contracts increasingly contain AI clauses specifying whether a client may train models on recorded sessions. But contracts only bind the parties who signed them; they do nothing against a stranger who scrapes your podcast appearances and sells a clone online.
Practical Steps Voice Actors Can Take Today
Waiting for federal legislation is not a strategy. Voice actors can take concrete steps now to reduce exposure and strengthen future claims. First, audit where your voice exists online: every podcast appearance, YouTube video, audiobook sample, and demo reel is potential training data. You cannot remove everything, but knowing your footprint helps you detect misuse faster.
Second, put AI language into every contract. Even outside union coverage, freelancers can negotiate clauses stating that recordings may be used only for the specified project, that no machine-learning training or voice synthesis is permitted without separate written authorization, and that any such authorization carries additional fees. Clients increasingly accept these clauses because they want clean provenance too.
Third, register what you can. While voices themselves are not copyrightable, specific performances fixed in recordings are. Copyright registration of key recordings creates a public record and enables statutory damages if the recording itself is copied. Trademark protection, as the Taylor Swift filings illustrate, is an emerging option for distinctive voices tied to a brand, though it remains untested territory for most performers.
Fourth, monitor actively. Services exist that scan for cloned voices, and simple alerts on your name plus terms like "AI voice" catch many cases early. Detection matters because remedies shrink when infringement goes unnoticed for years.
Comparing Your Protection Options
| Protection Route | What It Covers | Strengths | Limitations |
|---|---|---|---|
| NO FAKES Act (if passed) | Federal right in voice/likeness, nationwide | Uniform standard, statutory damages, post-mortem rights | Not yet law; timing uncertain |
| State right-of-publicity laws | Varies by state; TN/CA strongest | Available now, no waiting | Inconsistent; weak or absent in many states |
| Contract AI clauses | Uses by contracting parties | Enforceable immediately, negotiable fees | Doesn't bind third parties |
| Copyright registration | Fixed recorded performances | Statutory damages up to $150,000 willful | Doesn't protect the voice itself |
| Union agreements (SAG-AFTRA) | Covered productions | Industry-wide consent standards | Only covers union work |
Common Mistakes Voice Actors Make
The most frequent mistake is assuming existing law already covers voice cloning comprehensively. It does not. Many performers discover only after an incident that their state lacks a usable claim, or that copyright doesn't reach the conduct at all. Another mistake is signing boilerplate contracts without reading the AI provisions — some current agreements grant sweeping rights to synthesize performances from session recordings, effectively licensing a clone for a one-time session fee.
A third error is overreacting in the other direction: refusing all AI-adjacent work outright. Consent-based voice licensing is becoming a legitimate revenue stream, and performers who understand the technology can negotiate paid replica deals on favorable terms while retaining control. Blanket refusal forfeits income without reducing risk, since bad actors don't ask permission anyway.
Finally, many performers neglect documentation. Keeping dated records of your recordings, contracts, and any discovered clones makes enforcement dramatically easier. An infringer's defense often collapses when the performer can show a clear timeline of original creation and unauthorized use.
When to Act and What It Costs
Act now, not after passage of the NO FAKES Act. Contract review costs nothing beyond time; an entertainment attorney reviewing or drafting AI clauses typically runs $200–$500 per hour, with simple clause additions to an existing agreement often billed at one to two hours. Copyright registration costs $45–$65 per work through the US Copyright Office's online system for single-author works. Monitoring services range from free manual Google Alerts to paid brand-protection platforms costing hundreds per month, though individual performers rarely need the top tier.
If you discover a clone of your voice, document it immediately with screenshots and archived links, send a takedown notice to the hosting platform (most major platforms have policies against non-consensual synthetic media), and consult an attorney before confronting the infringer directly. Under current law your leverage varies by jurisdiction, but platform policies often remove content faster than litigation ever would.
How Creators Can Work With Voices Ethically
This issue cuts both ways. Content creators who use voice tools — narration, dubbing, character work — face their own compliance obligations. The safe path is straightforward: use licensed voices from providers who obtained explicit consent from the underlying performers, keep records of those licenses, and never clone a real person's voice without written authorization even for parody adjacent to commercial use. Tools that enhance, clean, and process your own recorded audio sit comfortably on the legal side of the line; generating speech that imitates an identifiable person without consent does not, and the NO FAKES Act would make that distinction enforceable with federal damages.
Creators should also watch the consent-and-compensation market developing around ethical voice licensing. Performers are beginning to offer authorized voice models through managed marketplaces, creating a supply chain that respects the rights the NO FAKES Act seeks to codify. Building workflows around licensed voices now avoids painful retrofitting later.
The Road Ahead
As of August 2026, the NO FAKES Act has cleared Senate Judiciary Committee but faces an uncertain path through full congressional votes. Bipartisan sponsorship gives it better odds than most tech legislation, and industry support from both talent unions and parts of the AI sector (who want clear rules) improves the outlook. Still, bills die quietly every session, and voice actors should treat passage as possible rather than guaranteed.
Whatever happens federally, the direction of travel is unmistakable: consent before cloning, compensation for replication, and liability for unauthorized digital replicas. Performers who document their work, negotiate AI terms into contracts, and monitor for misuse will be positioned to benefit from stronger law when it arrives — and better protected in the meantime.