The ELVIS Act Isn't Just for Musicians

Photo by Mohamed Hamdi on Unsplash

Tennessee named their AI voice cloning law after Elvis Presley. I’ll give them that — it’s a better name than most AI legislation manages. The ELVIS Act — Ensuring Likeness, Voice, and Image Security Act — took effect July 1, 2024, and it’s substantially more important than the name suggests.

This isn’t a novelty law. It’s a meaningful expansion of right-of-publicity doctrine into AI territory, and it creates real legal exposure for anyone building voice cloning technology.

What Tennessee’s Existing Law Said

Tennessee already had a right-of-publicity statute. Under existing law, you couldn’t use someone’s name, photograph, or likeness for commercial purposes without their consent. That’s been the rule for decades.

The problem is that “voice” wasn’t clearly covered. And with the arrival of AI voice cloning — where you can feed a model twenty minutes of someone’s recordings and generate unlimited audio that sounds exactly like them — the old statutory language wasn’t keeping up.

The ELVIS Act fixes that. It explicitly adds voice to the list of protected attributes.

What the Law Does

The core of the statute is simple: no person may use artificial intelligence to produce a simulation of an individual’s voice for a commercial purpose without authorization.

“Commercial purpose” is defined broadly — it includes advertising, selling merchandise, soliciting purchases, or fundraising. It also includes using the simulated voice for entertainment purposes when done for profit.

The protected party doesn’t have to be a celebrity. The law covers any individual. Elvis’s estate benefits. So does the random person whose voice got scraped from a podcast without their knowledge.

Violations create a civil cause of action. The plaintiff can get:

• Injunctive relief — a court order to stop the conduct
• Actual damages or a $10,000 statutory minimum per violation, whichever is greater
• Profits the defendant made from the unauthorized use
• Attorney’s fees if the violation was willful

That attorney’s fees provision is the one that will make plaintiff’s lawyers excited. It turns this into a viable case even when the actual damages are modest.

Who This Actually Hits

The obvious targets are voice cloning services being used to generate fake audio of real people. But the law’s reach is broader than that.

If you’re building:

• Text-to-speech tools that allow users to clone arbitrary voices from audio samples
• AI dubbing or localization products that replace a speaker’s voice with a simulated version
• Synthetic media platforms where users upload real people’s audio to generate new content
• AI music tools that recreate a specific artist’s vocal style

…then the ELVIS Act creates potential liability for you — or your users — depending on how your product is designed.

The “or your users” part is the annoying part. The statute targets whoever “uses artificial intelligence to produce the simulation.” That could be your platform or the end user, and courts haven’t yet settled where platform liability ends and user liability begins in this context.

The Consent Question

The law requires “authorization” from the individual whose voice is being simulated. It doesn’t specify what form that authorization takes.

This means your standard terms of service language is going to matter a lot. If a user uploads their own voice to your platform and generates synthetic audio, you need their consent documented. If a user uploads someone else’s voice, your platform needs a process for verifying authorization — or you need to not allow that use case.

The music industry has been dealing with this most acutely. Independent artists are getting their vocal performances scraped to train models without their knowledge. The ELVIS Act gives them a Tennessee-law cause of action to fight back. The interesting litigation question is whether federal copyright law preempts some or all of these claims — that fight hasn’t resolved yet.

What Compliance Looks Like

If you’re operating a voice AI product:

1. Audit your use cases. Distinguish between users synthesizing their own voice and users trying to replicate someone else’s.
2. Document authorization. For any commercial voice synthesis involving a real person’s voice as a source, you need a clear consent trail.
3. Review your ToS. Make it explicit that users can’t use your platform to create unauthorized voice simulations of others. This won’t give you full immunity but it shifts the liability.
4. Talk to your Tennessee users. If you have enterprise customers in Tennessee using voice AI tools in customer service or advertising contexts, they need to understand this law applies to them.

The law extends to deceased individuals as well — if someone died after January 1, 1997, their voice rights survive for ten years post-death and are transferable to their estate. That’s directly the Elvis angle. His estate retains actionable rights over AI simulations of his voice.

The ELVIS Act is a reasonable law solving a real problem. It’s also a preview of where every state is heading — expect the right-of-publicity framework to expand to cover AI-generated likenesses nationally over the next few years. Tennessee just got there first.

You can find the original text of Tennessee Public Chapter 59 (ELVIS Act) on the Tennessee Secretary of State’s website.