Tennessee ELVIS Act and AI voice: a checklist for TTS projects

Tennessee Governor Bill Lee signed the ELVIS Act on 21 March 2024, according to the governor's news page, and the Act adds voice to the protections Tennessee gives a person's name, photograph and likeness. The page I read does not give an effective date or the statute text, so read the statute itself before relying on a detail.

What the page supports

The page is a signing announcement. It supports one claim: the law was signed and concerns voice. Anything finer, such as penalties or exceptions, needs the statute.

Consent checklist for AI voiceover

A checklist, not legal advice.

• Use stock or generated voices from the TTS model's catalog, not a named person's imitated voice.
• If a real person's voice is involved, keep their written consent.
• Store the job id and the voice id of each audio file.
• Do not describe a generated voice as a real person.

Sume side

The Sume TTS Router docs list Cartesia models such as sonic-3.5, billed at list times 1.25 by character. The docs do not describe voice cloning in the pages I read, so a consent record applies to whichever voice id you choose.