A DTC brand's founder records a two-minute voice sample, hands it to an AI production team, and asks for 50 ad variations reading different scripts in his own voice by the end of the week. That's a legitimate, fast-growing production workflow, right up until someone asks to clone a hired spokesperson's voice, or worse, a competitor's brand ambassador, without a written consent record on file. The technology doesn't distinguish between the two use cases. The law very much does.
Is voice cloning itself illegal, or only a specific way of using it?
Cloning a voice is not illegal anywhere the technology is sold. What's regulated is unauthorized commercial use of a specific, identifiable person's voice. In the US, about 30 states have a right of publicity law protecting the commercial use of a person's name, image, and voice, with California's statute the most frequently cited, and the FTC has made clear that deceptive uses of AI voice cloning fall squarely within its existing consumer-protection authority, there's no AI carve-out from laws already on the books.
The practical rule for a marketing team: cloning your own voice, or a voice you have documented written consent to use commercially, is safe. Cloning anyone else's voice for an ad without that paperwork is a liability regardless of how good the output sounds.
What did Tennessee's ELVIS Act actually change, and does it matter outside Tennessee?
Signed into law on March 21, 2024, the Ensuring Likeness Voice and Image Security (ELVIS) Act was the first US state law to explicitly protect an individual's voice as a property right against AI-generated replicas, expanding Tennessee's existing right of publicity statute specifically to cover voice cloning. It requires clear, written permission before using someone's real or AI-simulated voice for commercial purposes or advertising, with both civil and criminal exposure for violations.
It's a Tennessee statute, but it matters everywhere an ad might run or target Tennessee residents, and it set the template other states have been following since. Treat it as the current floor for voice-consent practice, not a regional exception.
What does the EU AI Act require before a cloned voice goes into a European ad campaign?
Article 50 of the EU AI Act, applying from August 2, 2026, requires disclosure whenever audio content is AI-generated or manipulated to resemble an existing person, the same deepfake-labeling requirement that applies to synthetic video and images. We've covered the full labeling and marking requirements under Article 50 in detail, the short version for voice specifically is that an audible disclosure or clear labeling is expected wherever a listener could otherwise mistake the cloned voice for the real person's own recording.
That sits on top of, not instead of, the consent requirement. A campaign can satisfy Article 50's disclosure rule and still be a right-of-publicity violation if the underlying voice was cloned without permission.
What does the FTC actually require for AI voice ads running in the US?
The FTC's March 2025 staff guidance on AI in advertising sets three core principles, transparency about AI involvement chief among them: if AI generated or substantially modified the ad content, that has to be disclosed to consumers, and an AI-generated endorsement or testimonial needs the AI involvement clearly stated. We go deeper on the mechanics of US disclosure requirements in our breakdown of the FTC's AI-generated advertising rules.
The Commission has also finalized an Impersonation Rule specifically giving it more direct enforcement tools against deceptive voice cloning, on top of its general deceptive-practices authority. For a marketing team, that means a compliant disclosure doesn't just avoid a complaint, it removes the ambiguity a plaintiff's attorney needs to build a right-of-publicity case in the first place.
How AIBOOTSTRAPPER helps
We built AudioBolo, an AI-powered audio platform, end to end for a founder who needed a working product shipped in weeks without compromising production quality, the same discipline applies to any AI voice production pipeline: get the underlying architecture and consent workflow right before scale, not after a legal notice forces a rebuild. For AI avatar, UGC, and voice production work specifically, we build the consent and disclosure step into the workflow from day one rather than bolting it on after a campaign is already live.
If your team is scaling AI voice or avatar content for ads and isn't sure your consent paperwork or disclosure practice would hold up, book a call before the next campaign ships, not after someone asks where the consent form is.
Want this done for you?
Book a free strategy call and we'll show you how to build and market your business with AI.