A Tennessee misdemeanor now carries up to 11 months and 29 days in jail, a $2,500 fine, and treble damages in civil court. The trigger: cloning someone's voice with AI, without their permission.
If you are a founder setting up a US entity to build anything touching voice, video, or likeness, a dubbing tool, a voice assistant, a synthetic avatar, an AI cover-song app, this law reaches you. Even if your company is registered in Delaware or Wyoming. Even if you have never set foot in Tennessee.
What the ELVIS Act does
Tennessee's Ensuring Likeness, Voice, and Image Security Act took effect on 1 July 2024. Governor Bill Lee signed it after country artists pushed back on AI tools cloning their voices without consent. It updates Tennessee's existing right-of-publicity law, the one protecting name, image, and likeness, and adds a new category: voice.
The definition is broad. It covers any sound "readily identifiable and attributable to a particular individual," whether that is a real recording or an AI simulation. You do not need to copy someone's exact words. A close-enough clone is enough to trigger a claim.
Why this is not just a music industry story
Building in one of these categories? Send me one message about what you're building and I'll name the laws you're inside — which documents they demand, and what covers you. Free, usually within a day or two. Start here.
— Santosh
Most publicity laws stop at the person who publishes the fake. The ELVIS Act goes further. It creates a separate claim against anyone who builds or distributes a tool whose "primary purpose or function" is producing an unauthorized voice or likeness replica.
Read that again. The liability does not stop at the customer who misuses your product. It can reach the company that built it.
If you are building a generative AI product, voice cloning, face-swap, synthetic avatars, AI dubbing, or even a deepfake-detection tool that also generates samples, this is the clause that matters most. Marketing your product as "clone any voice" puts you in a different legal position than marketing it as "clone your own voice, with verified consent." The wording of your landing page is now a legal decision, not just a marketing one.
Voice cloning was already on federal enforcers' desks before Tennessee acted. In early 2024, after an AI-cloned voice of President Biden was used in robocalls before the New Hampshire primary, the FCC ruled that AI-generated voices in robocalls are illegal under the TCPA and moved against the consultant behind the calls. The FTC finalized its rule on government and business impersonation the same year. The ELVIS Act adds a state-law claim on top of that federal pressure, and it is the one that reaches tool builders directly.
The penalties are not symbolic
A violation is a Class A misdemeanor: up to 11 months and 29 days in jail, plus a fine up to $2,500. On the civil side, courts can award injunctive relief, actual damages, and treble damages for willful violations, with attorney's fees added if the affected person served in the armed forces.
Three separate parties can sue: the individual, anyone holding rights to their likeness, and in some cases a licensee. This law was written to be enforced, not to sit on the books.
Tennessee is the start, not the whole story
Eleven states now restrict unauthorized voice cloning: Tennessee, California, Illinois, Indiana, Nevada, Montana, New Hampshire, New Jersey, New York, Pennsylvania, and Washington. Six of them, Tennessee, California, New York, Washington, New Jersey, and Montana, go further and also cover deepfakes tied to elections and sexual content.
Several ELVIS Act provisions have already shown up in proposed federal legislation. If you are building for the US market, treat this as the direction the law is heading, not a one-state exception you can route around.
Where you incorporate does not protect you
This is the part founders get wrong most often. A Wyoming LLC or Delaware C-Corp selling an AI voice product to US customers is still exposed if its tool clones a Tennessee resident's voice without consent.
Your state of formation affects taxes, banking, and how investors see your cap table. It does nothing to shield you from a publicity claim tied to where your users, or the person whose voice was cloned, are located.
Three things to fix before you scale
1. Build consent into the product, not just the Terms of Service. If your tool clones a voice, capture and log explicit consent from the voice owner at the point of use: who consented, to what, and when. A clickwrap buried in a ToS document will not hold up against a treble-damages claim.
2. Watch your own marketing language. "Clone any voice instantly" reads as exactly the "primary purpose" language the Act targets. "Clone your own voice, verified" does not. Audit your landing page, your app store listing, and your demo videos with this clause in mind.
3. Review this alongside your entity formation, not after your first user complaint. Founders default to thinking about company structure, EIN, and banking first, and treat product liability as a later problem. For anything touching AI voice or likeness, that order needs to flip. The consent architecture is harder to retrofit than the LLC was to form.
The one-line summary for each founder
Voice cloning and TTS: consent capture is now a product feature, and the tool builder can be sued directly.
AI dubbing and localization: your voice-actor and contributor agreements are the shield. Get them signed before the model trains.
Avatar and face apps: likeness is covered the same way voice is, and eleven states are watching.
Model and API providers: what customers do with your model can come back to you if your product's "primary purpose" reads wrong.
The ELVIS Act started as a music-industry protection bill in one state. For any founder building AI products for the US market, it is now a baseline compliance question, the same way sales tax nexus and data privacy became baseline questions a few years ago. The founders who build consent and verification into their product now will not be the ones explaining themselves in a Tennessee courtroom later.
Building anything that touches voice or likeness? Send me one message about what you're building, and I'll name the rules you're inside.