AI Voice Cloning Law: No Federal Rule, So Who Stops You

People keep asking whether AI voice cloning is legal in the United States, and keep getting answered as though a law exists. It does not. As of today there is no federal statute covering digital replicas of a person’s voice.

What binds you instead is a stack of three other things, all of which are in force right now and none of which require Congress to do anything.

Where the law actually is

In July 2024 the US Copyright Office published Part 1 of its AI report, dealing with digital replicas. Its conclusion was blunt: “we have concluded that a new law is needed. The speed, precision, and scale of AI-created digital replicas calls for prompt federal action.”

The Office wanted the law to cover everyone, not just the famous. Its recommendation on who should be protected says the statute “should cover all individuals, not just celebrities, public figures, or those whose identities have commercial value.”

It also drew a line that matters to producers. On what should trigger liability: “Liability should arise from the distribution or making available of an unauthorized digital replica, but not the act of creation alone.” Experimenting privately was not what they were aiming at.

That was 2024. Two and a bit years later, the legislation still has not happened.

The NO FAKES Act has not passed

The bill meant to fill the gap is the NO FAKES Act, S. 4591 in the 119th Congress. The Senate Judiciary Committee approved it by voice vote on 18 June 2026. It has not passed either chamber since. GovTrack’s record of the bill is the thing to check before anyone tells you otherwise, and we have linked it below rather than paraphrasing its status.

Two things follow. A bill that has cleared one committee is not law and does not bind you. And the 119th Congress ends on 3 January 2027, so if it does not get floor time before then it starts over.

We are reporting a legislative position that could change while this page is live. The date at the top is when we last checked it.

Layer one: the platform, which is already enforcing

This is the layer that will actually affect a release, and it is settled.

Spotify’s policy, from September 2025: “Vocal impersonation is only allowed in music on Spotify when the impersonated artist has authorized the usage.”

There is no ambiguity and no waiting for Congress. A convincing clone of a recognisable singer is a takedown, and the question of whether it was also unlawful never comes up. Platform policy moves faster than legislation and applies to you the moment you upload.

Layer two: the tool’s own contract

The terms you accepted when you signed up are enforceable against you regardless of what statute exists.

Suno’s terms, for instance, include a clause on voice models that has nothing to do with impersonation law at all: “In no event will you use the Output or your Voice Model to compete with Suno.” That is a contractual restriction on what you build, and it binds you because you agreed to it.

We went through the rest of Suno’s terms in detail in what you can actually sell from an AI tool. The general point applies to every tool in this space: read the voice-model clause specifically, because it is usually separate from the clause about ordinary output.

Layer three: state law, which copyright does not override

A common misreading is that federal copyright law somehow occupies this field. The Copyright Office addressed it directly, recommending it “should be clarified that section 114(b) of the Copyright Act does not preempt or conflict with laws restricting unauthorized voice digital replicas.”

It also recommended against full federal preemption if a new law arrives: “Given well-established state rights of publicity and privacy, the Office does not recommend full federal preemption. Federal law should provide a floor of consistent protection nationwide, with states continuing to be able to provide additional protections.”

So state rights of publicity and privacy are live, and the Office expects them to stay live even after a federal law passes. We have not surveyed individual state statutes and are not going to pretend otherwise. If you are doing this commercially, that survey is the homework, and it is the kind of question a lawyer should answer for your state rather than a website.

Style is a separate question

Producers often collapse two different things: making something that sounds like an artist, and making something that sounds like that artist’s actual voice.

The Copyright Office kept them apart. On style imitation it said it does “not, however, recommend including style in the coverage of new legislation at this time”, while recommending protection for realistic replicas.

A track in the manner of someone, with your own vocalist, sits in very different territory from a track with a clone of their voice on it. That distinction is doing a lot of work in this field and it is worth being precise about which one you are doing.

What this means in practice

Get permission before you clone a real person’s voice, not because a statute currently requires it everywhere, but because Spotify already requires it, the tool’s contract may restrict it, and state law may reach it. Three live constraints beat one hypothetical one.

Read the voice-model clause in your tool’s terms separately from the output clause. They are usually different and the voice one is usually stricter.

Keep private experiments private. The Copyright Office’s recommended standard was distribution, not creation, and that is a sensible line to hold even before it becomes law.

If you want the sound without the person, do it with style rather than with their voice. That is both the safer position and, on the Office’s analysis, the legally distinct one.

And watch this change. A bill that has cleared committee can move quickly. We will re-check and re-date this page, which is the only honest way to cover a question whose answer has a shelf life.

Sources

Each of these was opened and read on the date shown.

1. U.S. Copyright Office, Copyright and Artificial Intelligence, Part 1: Digital Replicas (July 2024), read 8 November 2026
2. Spotify, Spotify Strengthens AI Protections for Artists, Songwriters, and Producers (25 September 2025), read 7 November 2026
3. Suno, Terms of Service (revised 10 August 2026), read 7 November 2026
4. GovTrack, S. 4591 NO FAKES Act of 2026, bill status, read 8 November 2026

Found something out of date or wrong? Tell us. We will fix it and say that we did.