// AI NEWS · POLICY

California Will Require Ads to Disclose AI-Generated "Synthetic Performers" Starting in 2027

Governor Newsom signed SB 1050 on September 16, 2026. From January 1, 2027, an ad that prominently features a fabricated AI human must say so — but the law deliberately does not cover AI copies of real people, which sit under a separate consent regime.

2026-09-26 · by Moe Ameen

What happened

On September 16, 2026, California Governor Gavin Newsom signed SB 1050, a disclosure law aimed squarely at AI-generated people in advertising. Starting January 1, 2027, an advertisement that "prominently features" a "synthetic performer" must clearly and conspicuously disclose that fact. The required notice is short and plain — something substantially similar to "this performance features a synthetic performer" or "no human performer is depicted."

The law defines a synthetic performer as a digital figure, voice, or representation created in whole or in part with generative AI that produces the realistic impression of a human performance. The definition carries an important limit: the figure must NOT be recognizable as an identifiable real person. In other words, SB 1050 governs fabricated, invented humans — the AI "actor" who never existed — not deepfakes of a specific individual. "Prominently featured" is scoped too: the disclosure duty triggers when the synthetic performer appears in the foreground demonstrating the product or service, delivers voiceover or narration, or illustrates or reacts to the narration. Cartoon and clearly anthropomorphized characters are outside the definition, and there are carve-outs for advertisements promoting expressive works such as films, TV, and video games (where the synthetic performer's use matches its use in the work) and for cases where generative AI only handles language translation or accessibility features.

The consent question that most people mean when they say "AI likeness in advertising" — an ad that features an AI-generated version of a real person without clear permission — is handled by a different set of laws, not SB 1050. In California, using a digital replica of a real performer is already constrained by AB 2602 (effective January 1, 2025), which bars contract terms that let a company use a performer's AI replica without informed consent and representation, and AB 1836 (effective January 1, 2025), which requires estate consent to use a digital replica of a deceased performer. At the federal level, the FTC treats an undisclosed synthetic endorser as potentially deceptive, and the proposed NO FAKES Act — which would create a nationwide right against unauthorized AI digital replicas of a person's voice and likeness — remains pending in Congress. SB 1050 fills the remaining gap: the ad that uses a convincing person who is not real, where no individual's rights are triggered but a viewer could still be misled about whether a human was involved.

Enforcement runs through California's existing consumer-protection machinery. A violation is treated as a breach of the state's False Advertising Law and Unfair Competition Law, which opens the door to private enforcement and potential class-action exposure rather than a single regulator's discretion. Treat the exact statutory text and any implementing guidance as the controlling source before you build compliance around it; the practical shape, though, is clear — as of 2027, "was a real human in this ad?" becomes a question California expects advertisers to answer on the face of the creative.

Why it matters for creators

  • There are now two distinct exposure lanes for AI people in ads, and they need different fixes. A fabricated, invented performer needs a disclosure line (SB 1050). An AI copy of a real person needs documented consent (AB 2602, AB 1836, the FTC, and the pending NO FAKES Act). Solving one does nothing for the other.
  • Disclosure attaches to how prominently the synthetic performer is used, not merely whether AI touched the ad. Foreground demonstration, voiceover, and reacting to narration are the triggers — a background AI extra is treated differently from an AI spokesperson.
  • The penalty path is consumer-protection law, not a niche media rule. Because violations run through California's False Advertising and Unfair Competition statutes, the realistic risk is private litigation and class actions, which raises the stakes for any brand running paid creative into California.
  • State-by-state divergence is now the operating reality. California's disclosure rule sits alongside New York's advertising-transparency law, Hawaii's digital-imitation consent statute, and the EU AI Act's deepfake labeling duty — a multi-front compliance map for anyone advertising across regions.
  • The safest production posture is generating from an identity you own and can consent to. Content built on a persona you created sidesteps the real-person consent problem entirely and makes the synthetic-performer disclosure a one-line, deliberate addition rather than a scramble.

How to act on this with Kompozy

This law rewards a specific discipline: know exactly whose face and voice are in your ad, and be able to say so. That is the design premise of how [Kompozy](/) makes video and images. Its avatar and image output — [Persona Shorts](/glossary/persona-shorts), Persona HeyGen video, and face-locked [Persona Photos](/glossary/output-buckets) — is generated from an AI Influencer persona you set up and control, governed by a [Persona Brief](/glossary/persona-brief). You are never borrowing a celebrity's likeness or scraping a stranger's face, so the AB 2602 / AB 1836 / NO FAKES consent lane simply does not apply to your creative. If the on-screen figure is a synthetic persona you built, it is your identity to use — and if it is a real person on your team, you have their consent by construction. That is the clean side of both regimes, and it is where Kompozy puts you by default.

Where SB 1050 does bite — an ad whose spokesperson is a fabricated AI human — the fix is a disclosure line, and Kompozy is built to add it without friction. Every generated post is editable copy, and nothing ships unattended: [Autopilot](/glossary/autopilot) fans your creative across the eight social platforms plus blog and email, but each post can pass through a per-post review pipeline before it publishes, so stamping "this performance features a synthetic performer" onto the caption or on-screen text of an affected ad is a deliberate step, not an afterthought. One brand identity, one consented persona, one place to add the required notice, then distribution everywhere — the opposite of discovering after launch that the same AI-spokesperson creative went out across ten placements without a word about it. For the deeper legal and production playbook, see [unauthorized AI likeness ads: the consent line and how to produce likeness-based ads you can defend](/guides/unauthorized-ai-likeness-ads) and, on the disclosure-detection side, [where YouTube's AI disclosure and likeness rules stand in 2026](/news/youtube-ai-disclosure-likeness-rules-2026).

Quick takeaways

  • California Gov. Newsom signed SB 1050 on September 16, 2026; it takes effect January 1, 2027.
  • It requires a clear, conspicuous disclosure when an ad prominently features a "synthetic performer" — an AI-generated realistic human who is NOT an identifiable real person.
  • Accepted disclosure wording is substantially like "this performance features a synthetic performer" or "no human performer is depicted."
  • It does not cover deepfakes of real people; those fall under California AB 2602 / AB 1836, the FTC, and the pending federal NO FAKES Act.
  • Violations are enforced as False Advertising Law and Unfair Competition Law breaches, exposing advertisers to private litigation and class actions.

Frequently asked questions

What does California SB 1050 require?

Starting January 1, 2027, an advertisement that prominently features a "synthetic performer" — a realistic human figure, voice, or representation created with generative AI — must clearly and conspicuously disclose it, using wording substantially like "this performance features a synthetic performer" or "no human performer is depicted." Governor Newsom signed the bill on September 16, 2026.

Does SB 1050 cover AI deepfakes of real people in ads?

No. SB 1050 is limited to synthetic performers who are NOT recognizable as an identifiable real person — fabricated, invented humans. An ad that uses an AI-generated version of a specific real person without consent is governed by other laws: California AB 2602 (unconsented digital replicas of living performers) and AB 1836 (deceased performers), the FTC's authority over deceptive endorsements, and the proposed federal NO FAKES Act, which remains pending in Congress.

When does an ad have to disclose a synthetic performer?

The duty triggers when the synthetic performer is "prominently featured" — appearing in the foreground demonstrating the product or service, delivering voiceover or narration, or illustrating or reacting to the narration. There are carve-outs for advertisements promoting expressive works like films, TV, and video games (when the use matches the work) and for cases where AI only handles translation or accessibility. Cartoon and clearly anthropomorphized characters are outside the definition.

How can a creator or brand stay on the safe side of these rules?

Generate from an identity you own and can consent to. If your on-screen figure is a persona you created, the real-person consent laws do not apply, and adding the SB 1050 disclosure becomes a single editable line. Kompozy generates avatar video and face-locked images from an AI Influencer persona you control, governed by a Persona Brief, and routes every post through a review step before publishing — so you can add the required notice deliberately before the creative fans out across platforms.

Related news

← All AI news · Get started →