AI voice rights: NO FAKES hits an objection as state laws stand

A Senate attempt to pass the NO FAKES Act by unanimous consent met an objection on September 30, 2026. That attempt did not enact the proposal. The Congressional Record records both the request and Senator Ted Cruz's objection.

For singers, the distinction is practical. Proposed federal protection should not be reported as an existing nationwide remedy. At the same time, a delayed federal bill does not erase state laws or make every unauthorized voice clone lawful.

What protects a singer from AI voice cloning?

The answer depends on the law, the location, the use, and the permissions involved. A person's voice and a sound recording are different subjects. Owning an audio file does not answer every question about using it to simulate the performer.

The US Copyright Office's digital-replicas report examines gaps across existing protections. It recommended a federal right addressing unauthorized digital replicas. The recommendation covers people broadly, not only famous entertainers.

The report also recognizes the need to accommodate protected expression. That balance is part of the policy question. Supporting artist consent does not require describing all imitation, commentary, or parody as unlawful.

Our analysis: a release team should treat permission to make a voice replica as an explicit decision. Permission to record a session, edit that performance, and create new synthetic performances can have very different consequences for a singer.

What happened to NO FAKES in September 2026?

Senators Marsha Blackburn and Chris Coons sought passage of S. 4591. They described protections for voice and visual likeness, a notice-and-takedown process, and exceptions for expression including parody and commentary.

Cruz objected, arguing that the proposal needed stronger free-speech safeguards. His objection stopped that unanimous-consent request. It was not a final vote defeating the bill, and it was not a court ruling about a specific cloned song.

The bill record identifies the proposal as the NO FAKES Act of 2026. A bill title containing “Act” is not evidence of enactment. Readers need to distinguish a proposal, committee action, chamber passage, and a law.

This article reports the September 30 event checked on October 7. It does not predict whether negotiations will produce a later agreement or what a final version would contain.

Tennessee's ELVIS Act is a different legal event

Tennessee's governor signed the ELVIS Act on March 21, 2024. The state's announcement describes protection for voice alongside name, photograph, and likeness. It explicitly connects the change to unauthorized AI impersonation.

That is an enacted state measure, unlike a pending federal proposal. It is not a reason to assume identical rights, exceptions, procedures, and remedies in every state.

Development	Status established by the source	What readers should not infer
Tennessee ELVIS Act signing, March 2024	An enacted state law.	A uniform national rule.
Copyright Office report, July 2024	A recommendation for federal legislation.	Enactment by Congress.
NO FAKES Senate request, September 2026	An objection blocked unanimous consent.	Final defeat or passage of the bill.

The table separates legal events that can sound similar in a headline. For an artist considering action, the applicable law and facts still require individual attention.

Impact on users

Our analysis: a synthetic duet needs a clear account of who approved each voice. The permission should describe what can be generated, where it can appear, and whether the recording can be sold. A general statement that a collaborator “agrees to AI” leaves too much unresolved.

Consider a singer who approves a translated version of one chorus. That does not tell you whether the same model can create new songs, advertisements, or political statements. Those choices should be discussed before the voice becomes reusable production material.

Payment also needs a defined basis. A one-time session payment and an ongoing license for new synthetic performances are different arrangements. Accurate credits should explain the singer's actual contribution without suggesting that they performed words they never approved.

For a lyric video, the visual presentation can add to the impression of endorsement. A name, portrait, or featured-artist credit can make listeners believe the singer joined the release. Review the whole package, including the title and description, with the approved use in mind.

If somebody has copied your voice

Our practical recommendation is to preserve the original evidence before it changes. Save the public URL, date, account name, displayed credits, and a copy of the material where lawful. Record why you believe the voice identifies you and whether any permission was given.

Use the platform's relevant impersonation or synthetic-media reporting route. A copyright notice and a complaint about an unauthorized replica address different issues. Describe the right and conduct accurately instead of assuming one form fits every problem.

For a commercial release or repeated impersonation, qualified advice can identify the relevant state law, contract, or other claim. The existence of a federal proposal should not delay preserving evidence of a present problem.

Consent should survive the product launch

Our editorial position is that singers deserve meaningful control over their identity. A workable agreement explains approval, compensation, credit, permitted contexts, and what happens when the relationship ends. Those terms matter as much as the demo that sells the tool.

Our Suno voice and duet report addresses product controls. Our Spotify persona-badge report examines identification at the platform level. Neither a feature nor a badge substitutes for the person's permission.