What does the Tennessee ELVIS Act actually prohibit?

The ELVIS Act prohibits using AI to create a realistic reproduction of any living or deceased individual's voice or likeness for commercial purposes without written consent from the individual (or, if deceased, whoever holds their personal rights). It extends Tennessee's Personal Rights Protection Act to explicitly cover AI-generated voice cloning.

Does the ELVIS Act only apply to celebrities?

No. The ELVIS Act applies to any person's voice or likeness, not just public figures or recording artists. Cloning a private individual's voice without consent for commercial use violates the law the same as cloning a famous singer's voice. The name references the legislative inspiration, but protections extend to everyone.

What counts as a commercial purpose?

A commercial purpose includes using the AI-generated voice or likeness in advertising, marketing, entertainment products, or any monetized context. Internal testing of AI voice tools generally is not commercial use, but using a cloned voice in a podcast ad, product demo, or promotional video is.

What is the consent requirement?

Consent must be explicit and in writing. General terms-of-service acceptance or a vague recording release is insufficient. The document should identify the specific person, describe the AI uses covered, specify duration and exclusivity, and be signed by the individual or authorized representative. Standard talent releases typically do not cover AI reproduction.

What are the penalties for violating the ELVIS Act?

The law creates a civil cause of action. Violators can be sued for actual damages plus profits from the unauthorized use. Courts may grant injunctive relief and order seizure of infringing materials. For knowing violations involving armed forces members, the Act provides treble damages plus attorney's fees. Enforcement is through private lawsuit, not an administrative agency.

Does this apply to AI voice tools that use generic synthetic voices?

No. If a voice tool generates a voice not modeled on any specific real person, the ELVIS Act does not apply. The law targets reproduction of a specific identifiable individual's voice or likeness. The test is whether a reasonable listener would recognize the output as a reproduction of a specific real person.

Tennessee ELVIS Act: AI voice and likeness…

Tennessee ELVIS Act: AI voice and likeness compliance guide 2026

Tennessee's ELVIS Act (effective July 1, 2024) makes it illegal to use AI to clone a real person's voice or likeness without their written consent for commercial purposes. This guide covers who is affected, what consent you need, the civil remedies available (actual damages plus profits, injunctive relief, treble damages for knowing violations involving military personnel), and a decision tree for teams using AI voice tools.

TL;DR: Tennessee's ELVIS Act bars using AI to clone any real person's voice for commercial use without written consent. Civil remedies include actual damages, profits from the infringement, and injunctive relief. Generic synthetic voices not based on a specific person are outside the law's scope.

Tennessee enacted the Ensuring Likeness Voice and Image Security (ELVIS) Act with an effective date of July 1, 2024. The law fills a gap that had existed in personal rights protection: prior law covered photographs and visual likenesses but did not specifically address AI-generated voice cloning. The ELVIS Act extends those protections to voice and extends the mechanism of violation to cover AI-generated reproductions.

Two years into enforcement, teams building audio content, marketing campaigns, and AI-powered products need a clear understanding of when they need a consent license, what that license must cover, and what happens if they skip it.

What the law does

The ELVIS Act amends Tennessee's existing Personal Rights Protection Act (PRPA) to add protection for a person's voice and to make explicit that AI-generated reproductions of voice and likeness are covered.

Under the law, it is unlawful to use AI or similar technology to create a realistic reproduction of an individual's voice or likeness for commercial purposes without:

The individual's explicit written consent (or, for deceased individuals, consent from the person or entity holding their personal rights under Tennessee law), and
Any use that falls within the scope of what that consent actually covers.

A consent that covers a single podcast advertisement does not authorize using the same cloned voice in a billboard campaign or a streaming music release. Each distinct commercial use needs to be within the scope of the written consent.

Who this affects

The ELVIS Act is not limited to the music industry, though that is where most of the public attention has focused. Any organization that creates AI-generated audio content using real voices has potential exposure:

Advertising agencies and marketing teams using AI voice tools to recreate a spokesperson's voice for localized content
Podcast producers and content creators using AI to generate "interviews" with real people or to mimic a recognizable voice
Game studios and entertainment companies using AI voice synthesis to generate dialogue in the style of a real actor or performer
Corporate communications teams using AI to create video narrations in the voice of a company executive for internal or external distribution
AI product developers training or fine-tuning voice models on recordings of specific real individuals

Any organization outside Tennessee that creates AI voice content used commercially in Tennessee, or creates content that reproduces a Tennessee resident's voice, should also consider whether the law applies.

The key distinction: specific person versus generic AI voice

The most practically important question for teams using AI voice tools is whether the tool is reproducing a specific real person's voice or generating a novel synthetic voice.

Covered by the ELVIS Act: Using ElevenLabs, Murf, Resemble AI, or any similar tool to clone the voice of a specific named individual whose recordings were used as training or reference material. The test is whether a listener would recognize the output as that person's voice.

Not covered by the ELVIS Act: Using the same tools to generate a synthetic voice that sounds human but does not replicate any specific real person. Creating a fictional narrator voice with generic characteristics is outside the law's scope, as long as the AI is not reproducing a specific real person's vocal patterns.

The boundary can be blurry. If you ask an AI voice tool to generate "a voice similar to [famous person]" and the output is a close match, you are likely in covered territory even if you did not explicitly provide recordings of that person. Assess the output, not just the prompt.

Audio mixing equipment representing content production and compliance

Work through these questions before creating or publishing AI-generated audio content:

Question 1: Is this AI voice tool generating content that sounds like a specific real person?

No (it's a generic synthetic voice): ELVIS Act does not apply. Stop here.
Yes or unclear: Continue to Question 2.

Question 2: Is the content being used for a commercial purpose?

No (purely internal testing, no commercial distribution): ELVIS Act likely does not apply. Document your rationale.
Yes (advertising, marketing, paid content, product distribution): Continue to Question 3.

Question 3: Do you have written consent from the individual (or their authorized representative) that covers this specific use?

Yes: Verify consent scope matches the use. Proceed.
No: You need consent before proceeding. Do not publish without it.

Question 4 (if you cannot get consent): Can you use a different AI voice that does not replicate a specific real person?

Yes: Switch to a generic synthetic voice. Problem resolved.
No (the specific voice is essential to the project): Engage legal counsel. You may be looking at ELVIS Act liability.

Standard talent release forms and work-for-hire agreements often do not adequately address AI voice reproduction. Before using any recording as the basis for AI voice cloning, confirm your agreement covers:

Identification of the individual: Full name and a statement that consent applies to AI-generated reproductions of their voice and likeness.
Scope of permitted uses: List specific use cases (advertising, narration, in-app, etc.). A catch-all "all commercial purposes" is cleaner but may be harder to get talent to sign.
Duration: Is the consent perpetual or time-limited?
Exclusivity: Can other parties also license this person's AI-cloned voice, or does your license grant exclusive use?
Modification rights: Can you modify the AI-generated voice output (pitch, tone, language), or must it be used as generated?
Revocation terms: Under what circumstances can the individual withdraw consent?
Compensation: Specify any payment, royalty, or other consideration.

Work with your entertainment or IP counsel to create a standard AI voice consent addendum that attaches to your existing talent agreements. This is faster than renegotiating from scratch each time.

Penalties and enforcement

The ELVIS Act creates a private civil cause of action. The individual whose voice or likeness was reproduced without consent can sue and seek:

Actual damages: Economic harm (lost licensing fees the person could have charged, harm to their professional reputation, impact on their existing licensing relationships) plus non-economic harm.
Injunctive relief: A court order requiring the defendant to stop using the unauthorized content and remove it from distribution.
Actual damages: Economic harm (lost licensing fees, professional reputation damage, impact on existing licensing relationships) plus non-economic harm.
Profits: Any profits attributable to the unauthorized use not already captured in actual damages.
Treble damages + attorney's fees: Available when the defendant knowingly uses the voice or likeness of a member of the U.S. armed forces or Tennessee National Guard without authorization.
Attorney's fees: Courts may award the prevailing plaintiff's legal costs at their discretion.

There is no state agency that prosecutes violations. Enforcement is plaintiff-driven, meaning you will hear from a lawyer, not a regulator, when a violation is alleged. This is both better and worse than administrative enforcement: there is no compliance examination process, but there is also no opportunity to voluntarily disclose and negotiate a reduced penalty before a lawsuit is filed.

The ELVIS Act in relation to other AI content laws

Tennessee is not alone in regulating AI-generated likenesses. As of mid-2026, a growing number of states have enacted similar laws, and federal proposals are in circulation. The ELVIS Act sits alongside:

Washington State's AI likeness law: Similar protections covering AI-generated voice and visual likeness, with some differences in consent requirements. See Washington State AI likeness law June 2026 compliance.
EU AI Act Article 50: Requires disclosure (not consent) when AI is used to generate audio-visual content involving real people, including synthetic voice content. The obligation is transparency, not prohibition. See EU AI Act Article 50 watermarking and deepfake disclosure.
Take It Down Act: Federal law targeting non-consensual intimate images including AI-generated deepfakes. Narrower than ELVIS but covers overlapping territory for sexual content. See Take It Down Act compliance 2026.
California SB 942: Requires disclosure labeling on AI-generated content, including synthetic audio. See California SB 942 AI transparency act August 2026.

Organizations operating across multiple jurisdictions need to track all of these, not just the one that applies in their home state. A piece of AI-generated audio content distributed nationally may need to satisfy Tennessee consent requirements, Washington consent requirements, EU disclosure requirements, and California labeling requirements simultaneously.

Practical steps for teams using AI voice tools

If your organization creates any AI-generated audio or audio-visual content, take these steps now:

Audit your AI voice tool stack. List every tool used to generate voice content. For each tool, determine whether any outputs have been used that replicate a specific real person's voice.
Review existing consent coverage. For any projects that used a real person's voice as source material for AI cloning, confirm you have a valid written consent document covering the uses made.
Update your standard agreements. Add an AI voice reproduction clause to all talent and contractor agreements going forward.
Brief your creative and content teams. The people making decisions about AI voice tool use every day need to understand the line between generic synthetic voices (fine) and reproductions of specific real voices (needs consent).
Create an approval workflow. Any project involving AI voice content based on a real person's voice should route through legal review before production, not after.
Update your acceptable use policy. If your organization has an AI acceptable use policy (see AI acceptable use policy template small teams), add a section on voice and likeness reproduction rules.

For the broader AI voice and likeness law picture across states, see Washington State AI likeness law June 2026 compliance. For the related but distinct disclosure obligation covering fully AI-generated (not real-person) performers in ads, see New York AI synthetic performer ad disclosure law, effective June 9 2026. For EU disclosure requirements on AI-generated content, see EU AI Act Article 50 watermarking and deepfake disclosure. For federal non-consensual image law, see Take It Down Act compliance 2026. Track effective dates for AI laws at AI regulation deadline calendar 2026. For California synthetic media labeling rules, see California SB 942 AI transparency act August 2026. For your broader governance program, see AI governance checklist 2026.

One practical note for teams that create marketing and advertising content using AI voice tools: the consent requirement applies to the person whose voice is being replicated, not to the brand. If a talent agency owns the commercial booking rights to an individual's name and likeness, it does not follow that the agency can also consent to AI voice cloning on that person's behalf unless the talent agreement explicitly grants that right. Always confirm who has the authority to execute a voice reproduction consent agreement before committing to a production that depends on it.

Tennessee ELVIS Act: AI voice and likeness compliance guide 2026