California has enacted SB 1050, adding Business and Professions Code section 17610, under which, once it takes effect, creating and causing publication of an advertisement that prominently includes a generative-AI synthetic performer without a clear and conspicuous disclosure would be an unlawful advertising practice

Signed on 16 September, operative on 1 January 2027 by default

The chapter header on the chaptered text records both events on one day: approved by the Governor on 16 September 2026 and filed with the Secretary of State on 16 September 2026, as Chapter 246, Statutes of 2026. The official bill history lists the same two entries against 09/16/26. The leginfo page itself carries a Date Published stamp of 17 September 2026, 10:00 AM. That is when the page went up, not when the law was made.

Nothing in the act says when it takes effect. There is no urgency clause, no delayed operative date and no separate compliance date. The date comes from outside the bill. Article IV, section 8(c)(1) of the California Constitution provides that a statute enacted at a regular session goes into effect on January 1 next following a 90-day period from the date of enactment. Ninety days from 16 September 2026 runs into December, so the section becomes operative on 1 January 2027. That is a derivation from the constitutional default and we present it as one; the bill does not print the date.

Until then, section 17610 binds nobody. Every description in this piece of what the section requires is a description of what it would require from 1 January 2027.

Four definitions decide who is inside

Subdivision (b) is one sentence, and the sentence turns on defined terms. Under the new section, once it takes effect, a person who creates and causes to be published in an advertising medium an advertisement that prominently includes a synthetic performer would need a clear and conspicuous disclosure that the advertisement includes a synthetic performer. Read with the definitions in subdivision (a), that is narrower than a rule about AI in advertising.

Once section 17610 is operative on 1 January 2027, synthetic performer, as defined in paragraph (a)(6), would mean a digital figure, voice or representation created in whole or in part using generative artificial intelligence that creates the realistic impression of the audio, audiovisual or visual performance of a human performer who is not recognizable as any identifiable natural person. Two limits sit inside that. The figure has to create a realistic impression of a human performance in audio, audiovisual or visual form, so obviously nonrealistic visuals alone do not establish coverage; a realistic synthetic human voice over an animated figure would still need assessing against the words. And the performer has to be one not recognizable as any identifiable natural person, which puts a digital replica of a real, recognizable individual outside this definition. The section says nothing about replicas of real people, and we do not read it as covering them.

From 1 January 2027, prominently, in paragraph (a)(5), would have three limbs, any one of which is enough: the synthetic performer is in the foreground and demonstrating or illustrating the product or service; it is providing or voicing the on- or off-camera narration or commercial message; or it is illustrating or reacting to that narration or message. A synthetic figure in the background of a shot, doing none of those things, is not prominent under the statutory test.

Once the section is operative, advertisement, in paragraph (a)(1), would mean any audio, video or audiovisual message, statement, audiovisual recording, digital communication or other representation disseminated in any manner or by any means, including through online platforms, that is intended to induce, or is reasonably expected to induce, the purchase of goods or services, as described in section 17500. Advertising medium, in paragraph (a)(2)(A), lists broadcast stations, cable operators, multichannel video programming distributors, online platforms, streaming services, digital advertising networks, publishers and any other person or entity that distributes, displays, transmits or makes available an audio or audiovisual advertisement to consumers in this state. Subparagraph (B) then removes anyone not capable of removing, disabling access to or ceasing further dissemination of an advertisement. Whether a purely static image campaign, or a particular intermediary in the ad supply chain, sits inside those two definitions is a question the text leaves open, and we do not answer it here.

The disclosure wording is prescribed, within a tolerance

Once operative, subdivision (c) would not leave the label to the advertiser. A disclosure under the section would need to use wording substantially similar to "this performance features a synthetic performer" or "no human performer is depicted." Substantially similar is the standard and the two phrases are the anchors. Exact reproduction is not demanded by the words, and an unrelated AI label is not obviously within them either. Which departures still count as substantially similar is not something the text settles.

Clear and conspicuous is defined too, in paragraph (a)(3), and from 1 January 2027 it would mean a disclosure that is difficult to miss, easily understandable, and presented in a manner that a reasonable consumer would notice, read and comprehend, taking into account the medium, format and context in which the advertisement appears. That last clause makes the test medium-specific. What is conspicuous in a thirty-second broadcast spot and what is conspicuous in an audio-only ad are different questions, and the section gives no font size, duration or placement rule for either.

Five things the section says it does not do

Subdivision (d) is a list of construction rules, and two of them are exemptions; none of it is operative before 1 January 2027. From that date the section is not to be construed to restrict or prohibit the creation, distribution or exhibition of synthetic content, paragraph (d)(1). It does not regulate the expressive or informational content of an advertisement, except to require the factual disclosure in subdivision (b), paragraph (d)(2). It does not affect or limit any rights or remedies under section 17200 or 17500 or any other law concerning deceptive, unfair or misleading business practices, paragraph (d)(3). The first of those answers anyone reading this as a restriction on using synthetic performers: it is a labelling rule, and the text says so.

Once operative, paragraph (d)(4) would take advertisements for expressive works out of the section, the list running to motion pictures, television programs, streaming content, documentaries, video games or other similar audio or audiovisual works, provided that the use of a synthetic performer in the advertisement or promotional material is consistent with its use in the expressive work. A trailer that shows a synthetic character as it appears in the film would not need the label. The proviso does real work: a synthetic performer used in the promotion in a way that departs from its use in the work is not within the exemption's words.

Once operative, paragraph (d)(5) would exempt an advertisement in which the use of generative artificial intelligence solely involves the language translation of a human performer or other accessibility features. Solely is the operative word. An ad that uses generative AI to translate a human actor's lines, and for nothing else, would be outside the section.

Platforms act on a served court order, not on a complaint

A reader expecting a takedown regime will find a narrower and more procedural one, and none of it operates before 1 January 2027. From that date, subdivision (e)(1) would prohibit an advertising medium from transmitting, distributing, displaying, airing or otherwise making available an advertisement containing a synthetic performer only if both of two conditions are met: a court of competent jurisdiction has issued an order finding that the advertisement violates the chapter or enjoining the creator from publishing it, the text giving an injunction under section 527 of the Code of Civil Procedure as an example; and the advertising medium is served with the order and information reasonably sufficient to identify the advertisement.

Once operative, on receipt of such an order, subdivision (e)(2) would require the medium, as soon as commercially reasonable and technically feasible, to do two things: remove, disable access to or cease further dissemination of the advertisement on the medium in this state, and cease accepting payment for further dissemination of it. There is no fixed number of hours or days. There is no duty to screen inbound creative, no duty to act on a consumer complaint, and no duty triggered by the medium's own knowledge that a label is missing. The trigger is a court order, served.

Enforcement borrows the existing false advertising machinery

Once operative, subdivision (f) would make a violation of the section a violation of section 17500 and provides that it may be enforced pursuant to Chapter 5 (commencing with section 17200) of Part 2. The section creates no penalty schedule of its own and names no private right of action of its own; what remedies follow is a question about sections 17200 and 17500, which we did not read for this piece. The Legislative Counsel's Digest printed with the bill states that existing law makes false or misleading advertising practices a misdemeanor and that, by expanding the scope of a crime, the bill imposes a state-mandated local program. SEC. 3 of the act then disclaims reimbursement on the ground that the act creates a new crime or infraction or changes the definition of a crime. We report those as the Digest's and the act's own characterisations. Subdivision (g) makes the requirements severable.

SECTION 1 records the Legislature's findings: that consumers have a substantial interest in knowing when an advertisement uses a synthetic performer in a manner that could reasonably be interpreted as a human performance, that a factual disclosure is necessary to prevent deceptive or misleading advertising practices, and that the disclosure is intended as a purely factual commercial disclosure consistent with constitutional protections for commercial speech. That last finding reads, to us, as drafted with commercial-speech doctrine in view. The act says nothing further about it.

What we did not verify

What we opened: the chaptered text of SB 1050 as served by California Legislative Information at bill_id 202520260SB1050, version 20250SB105093CHP, read end to end, including the chapter header, the Legislative Counsel's Digest, SECTION 1, SEC. 2 (section 17610, subdivisions (a) to (g)) and SEC. 3; the official leginfo bill history entries for 09/16/26; and the text of article IV, section 8 of the California Constitution as served by leginfo.

What we did not open: the Governor's 16 September 2026 press release, which we treat as a summary and not as the statute; Business and Professions Code sections 17200 and 17500 and the remedies that attach to them; Code of Civil Procedure section 527; the committee analyses, floor analyses and earlier amended versions of the bill; and New York's S8420-A or any other jurisdiction's synthetic performer statute.

What we refuse to claim: we do not say this is the first such law in the United States, or that it is identical to New York's, because we did not compare the texts. We do not say every AI-generated advertisement, or any digital replica of a real person, is covered, because the definitions in subdivision (a) say otherwise. We do not say static image campaigns or any particular intermediary are covered, because the interaction of the advertisement and advertising medium definitions does not settle that. We do not say platforms have to screen advertisements or remove them on complaint, because subdivision (e) requires a served court order. We do not say the section is operative now, because it is not. We do not say the act bans synthetic performers, prohibits replacing human performers, or creates any consent, compensation or hiring right, because the enacted text does none of those things, whatever the press release says. We give no fine, damages figure or enforcement outcome, because the section supplies none and we did not read the cross-referenced law. We do not say any wording other than the two statutory phrases satisfies subdivision (c).

Informational analysis for working professionals, not legal advice. Confirm how any rule applies to your situation with qualified counsel.

Related briefings

• New York S8420-A: synthetic performer advertising disclosure
• UK ASA monitoring of AI advertising disclosure
• AI disclosure rules tracker
• North America AI regulation tracker
• AI Regulation News hub