Can I use an AI voice that sounds like a real artist?

What are the current rules?

State law is the live enforcement layer. Tennessee's ELVIS Act (in force since 2024) explicitly covers voice as a protected property right and reaches AI simulation. Most other states have right-of-publicity laws that predate AI but apply to commercial voice imitation; scope varies by state.

Federal law is coming but not here. The NO FAKES Act of 2026 (S.4591) — a federal property right in voice and visual likeness — was reported by the Senate Judiciary Committee in June 2026 and sits on the Senate calendar. It has not passed either chamber. Earlier versions (2024, 2025) died or stalled; each is a separate bill, and our records track all of them by number.

Platforms enforce faster than courts. Spotify's current artist policy says it removes unauthorized music that replicates another artist's voice, including releases presented as an “AI version”; some parody uses may be exceptions. Other platforms have their own impersonation and deception rules.

Which uses are exposed and which are safe?

Using an AI voice that is a famous artist's voice for commercial release without consent: high exposure, and settlements in this space (disputed-likeness works are in our Works records) show rights holders act on it. Using generic synthetic vocals nobody would identify as a real person: none of the above applies. Parody, commentary and other contexts have their own First Amendment dimensions that no AI-specific court decision has mapped yet — that genuinely remains open.