The AI voice laws founders name do not reach a catalogue narrator, and one nobody names does

The right of publicity hooks on identifiability, and a catalogue narrator has none

Tennessee's Ensuring Likeness, Voice, and Image Security Act was enacted on 21 March 2024 and took effect on 1 July 2024. It extended the state's right of publicity to a simulation of a voice, and it reached whoever distributes a tool whose primary purpose is producing a particular, identifiable individual's voice, knowing the use was not authorised. The definition below is quoted from Latham and Watkins client alert 3244, dated 8 April 2024: Tennessee's own bill server refused the connection and the Secretary of State's chapter PDF answered 403. Voice means a sound in a medium that is readily identifiable and attributable to a particular individual, regardless of whether the sound contains the actual voice or a simulation of the voice of the individual. Two words carry that statute: identifiable, and particular.

California is the same shape and much older. Deciding Midler v. Ford Motor Co., 849 F.2d 460, in 1988, the Ninth Circuit quoted the state statute's list of name, voice, signature, photograph or likeness, and wrote that a voice is as distinctive and personal as a face. The tort it recognised needs a distinctive voice of a professional singer, widely known, deliberately imitated to sell a product. Distinctive, and widely known: the hook is in the holding twice.

Lehrman v. Lovo, decided in the Southern District of New York on 10 July 2025, shows what that hook costs a plaintiff. The court asked after the quality and quantity of the identifiable characteristics displayed, and let the New York Civil Rights Law claims under sections 50 and 51 proceed. Two named voice actors got there by showing their own voices were recognisable in the product, partly out of the defendant's own marketing, which called the clones practically indistinguishable from the real voice. A narrator who is nobody in particular cannot build that showing for anybody.

A right of publicity claim asks whose voice it is, and never asks how the voice was made.

The FCC ruling is about telephone calls, and its own press release is why nobody knows that

Declaratory Ruling FCC 24-17 was adopted on 2 February 2024 and released on 8 February. Paragraph 2 is the whole of it: the Commission confirms that the TCPA's restrictions on the use of artificial or prerecorded voice encompass current AI technologies that generate human voices, so calls using them require the prior express consent of the called party. Every operative sentence in the document is anchored to a call. So is the statute behind it, 47 U.S.C. 227(b)(1)(B), which makes it unlawful to initiate any telephone call to any residential telephone line using an artificial or prerecorded voice.

The press release that went out with it was headlined FCC MAKES AI-GENERATED VOICES IN ROBOCALLS ILLEGAL, and a few paragraphs down the same page says the ruling ensures AI-generated voices in calls are held to those same standards as other artificial voice calls. Both sentences sit in one document. Only the first one travelled. Wiley wrote in February 2024 that the Declaratory Ruling does not go that far. The correction landed inside the month and the headline still won.

The exposure is real and it is confined to calls: a private right of action at $500 in damages per violation, which a court may treble to $1,500 where the violation was willful or knowing, and per violation means per call because the prohibition is framed per call. A video file on a landing page places no calls. The mistake this produces is the expensive kind, a team abandoning synthetic narration on the strength of a telemarketing rule.

New York's statute triggers on a performer who is nobody in particular

Section 396-b defines a synthetic performer as a digitally created asset created, reproduced, or modified by computer, using generative artificial intelligence or a software algorithm, that is intended to create the impression that the asset is engaging in an audiovisual and/or visual performance of a human performer who is not recognizable as any identifiable natural performer. Where the right of publicity needs a particular identifiable individual, New York needs the absence of one. Bill S8420A became Chapter 617 of 2025 on 11 December, and the act's own clause set it running on the one hundred eightieth day after it shall have become a law, which lands on 9 June 2026.

Then the exemptions, and the first of them names audio. The section does not apply to audio advertisements, or where the use of artificial intelligence solely involves the language translation of a human performer. It does not apply to advertisements for expressive works, and nothing in it applies to the medium: television networks, streaming services, cable systems, billboards, transit advertisements.

Which leaves the case a product like this one actually produces. A synthetic voice over motion graphics is not an audio advertisement, and no synthetic performer is visually performing anything either. Honigman reads the audio exemption as covering synthetic narrators that exist without visual components, and says the law does not specifically address whether a realistic AI voice accompanying human visuals requires disclosure. No guidance, opinion or enforcement action on section 396-b settles it.

Reed Smith's advice is the one worth taking: there is no private right of action but the state can seek significant penalties, so err on the side of adding a disclosure if the ad could reach New York consumers. The statute does not define conspicuous, which the same firm reads as leaving that to market practice.

Tennessee needs a particular identifiable individual. New York needs the absence of one.

The FTC's own example says an unfamiliar announcer is not an endorser

The endorsement guides at 16 CFR Part 255 are where an advertiser expects to get caught, and the 2023 final version answers the narrator question in an example. Section 255.0(b) defines an endorsement as a message consumers are likely to believe reflects the opinions, beliefs, findings, or experiences of a party other than the sponsoring advertiser. Example 3 then puts an announcer unfamiliar to consumers except as a spokesperson for the advertising drug company in front of them, praising the product, not purporting to speak from personal experience but on behalf of the company. The Commission's conclusion: the announcer's statements would not be considered an endorsement.

A catalogue narrator reading a script the founder approved is Example 3. Not a person the audience knows, claiming no experience, speaking for the advertiser. What the 2023 revision did add reaches the other way, because the Commission said its new definitional language is intended also to encompass the writers of fake reviews and non-existent entities that purport to give endorsements. Invent a satisfied customer and the guides arrive. Hire a reader and they do not.

The sentence in Part 255 that does bear on a video is 255.0(f): where a representation is made through both visual and audible means, the disclosure should be made in the communication's visual and audible portions. That is the design rule for any disclosure a video carries, whichever statute asked for it.

Europe and California put the duty on the vendor, and California moved its start date

The European Commission's own FAQ gives the date: Article 50 of the AI Act applies as from 2 August 2026, with a limited grace period to 2 December 2026 for marking obligations on systems already on the market. Article 50(2) is the provider's, to ensure that the outputs of the AI system are marked in a machine-readable format and detectable as artificially generated or manipulated. Article 50(4) is the deployer's, and only for deep fake content, which sends the question to Article 3, where the same hook turns up. A deep fake there is AI-generated or manipulated image, audio or video content that resembles existing persons, objects, places, entities or events and would falsely appear to a person to be authentic or truthful. A narrator who resembles no existing person is not within it, so what is left is the marking duty under 50(2), which belongs to the provider rather than to the business publishing the advertisement. Article 99(4) prices a transparency breach at up to EUR 15 000 000 or 3 per cent of total worldwide annual turnover, whichever is higher.

California allocates it the same way, and has already moved once. SB 942 puts the duty on a covered provider, a person producing a publicly accessible generative artificial intelligence system with over 1,000,000 monthly visitors or users, at $5,000 per violation with each day of noncompliance a separate one. It set the chapter running on 1 January 2026. AB 853 replaced that clause: this chapter shall become operative on August 2, 2026. A summary written before that amendment carries the old date. A business publishing a video is not one of its addressees.

What to do about it

1. Read New York General Business Law 396-b before asking counsel about the ELVIS Act, and decide in advance whether an advertisement that could reach New York carries a synthetic performer disclosure.
2. Record which narrator was licensed and from where on every run, because Tennessee and California both hook on whose voice it is rather than on how it was made.
3. Put any disclosure in both the picture and the sound, which is what 16 CFR 255.0(f) asks of an audiovisual advertisement and what section 396-b leaves undefined.