AI Regulations in the US: What Brands and Creators Need to Know (2026)

Last updated: October 2026

As of October 2026, AI regulations in the US for marketing come from two layers: the FTC's existing rules on endorsements, disclosures and fake reviews, and a growing set of state laws on AI-generated performers and voice or likeness cloning.

There is no single federal AI law for advertising yet (a December 2025 executive order still calls for Congress to set a national standard), so a brand running AI-assisted ads or creator content has to follow both layers, plus each platform's own AI labels.

This guide is for brands and agencies using AI avatars, AI voiceovers, AI-edited UGC or AI-written scripts in creator campaigns, and for creators asked to license an AI version of themselves. It is general information, not legal advice. These rules change fast, so talk to a lawyer about specific campaigns.

AI regulations in the US: the quick answer by use case

Sort each AI use into one of four buckets. Each bucket has a different rule.

• An invented AI person in your ad (avatar, AI influencer, synthetic spokesperson): allowed, but New York requires a label now and California will from January 1, 2027. The FTC's Endorsement Guides also cover virtual influencers.
• An AI copy of a real person's face or voice: you need that person's permission. Tennessee and Hawaii give people legal rights over their voice and likeness, and California and Illinois limit vague contract clauses that hand over digital replica rights.
• AI-written reviews or testimonials: fake ones are banned under the FTC's Consumer Reviews and Testimonials Rule.
• AI edits or AI scripts on real creator content: the state label laws usually don't apply, but the FTC paid-partnership disclosure still does, and Meta, TikTok and YouTube may require or add an AI label when the result looks realistic.

Working with creators on AI-assisted content? On Jem Social you can brief creators, sign contracts, review drafts and approve them before payment leaves escrow, so disclosures and consents get checked before anything goes live. Start a 7-day free trial.

The FTC rules that apply even when AI made the content

The FTC has no AI-specific advertising rule. Its existing rules cover AI content the same way they cover a human creator.

Disclose the paid relationship

If you pay a creator or send free product, the post needs a clear disclosure. The FTC's Disclosures 101 for Social Media Influencers says to put it in the video itself, not only in the description, and not to assume a platform's built-in tool is good enough. A sponsored TikTok with an AI voiceover still needs a clear "ad" or "sponsored" disclosure.

Virtual influencers are covered too

When the FTC updated its Endorsement Guides in June 2023, it changed the definition of endorsements to clarify how it covers virtual influencers (FTC). Treat an AI character that promotes your product like a human endorser: true, backed-up claims and a clear brand connection.

An AI "customer" can't pose as a real one

The Endorsement Guides say ads that present endorsements from "actual consumers" should use actual consumers, or clearly disclose that the people shown are not (16 CFR 255.2(c)). An AI avatar playing a happy customer sits in that zone.

In its answers to business questions about the reviews rule, the FTC says there is "no blanket prohibition on the use of AI-generated avatars in marketing," but an avatar testimonial breaks the rule if the underlying testimonial is fake. So does an unauthorized celebrity avatar that viewers would take as a real endorsement.

No fake reviews, whether AI or a person wrote them

The Consumer Reviews and Testimonials Rule took effect October 21, 2024 and is in force as of October 2026 (FTC). It bans creating, selling or buying reviews and testimonials that misrepresent that they come from someone who does not exist, "such as AI-generated fake reviews," or from someone with no real experience of the product (FTC).

In December 2025 the FTC warned 10 companies about the rule and cited civil penalties of up to $53,088 per violation (FTC), an amount it confirmed stays the same for 2026 (Federal Register).

If you use AI to draft creator scripts, note this line from the Q&A: a business shouldn't provide testimonial text "without a reasonable basis to conclude that the text is truthful about the testimonialist's experience." Script the hook and product facts, and let creators describe their own experience.

AI avatars and synthetic performers: the state label laws

Two states have passed laws requiring a visible label when an ad uses an AI-generated person who isn't a real, identifiable individual. AI copies of real people fall under the likeness laws below.

New York: in effect since June 9, 2026

New York's law (S.8420-A) was signed December 11, 2025 (Governor's office) and its disclosure duty has applied since June 9, 2026, 180 days after signing. It is in effect as of October 2026. A business that produces or creates an ad must "conspicuously disclose" that a synthetic performer is in it, when it has actual knowledge.

A synthetic performer is an AI-made asset that looks like a human performer but isn't recognizable as any real person. Penalties are $1,000 for a first violation and $5,000 after that. Audio-only ads and translation-only AI are exempt, and the outlets running the ad are not covered.

California: signed September 16, 2026, starts January 1, 2027

California's SB 1050 was signed September 16, 2026. As a regular-session law, it takes effect January 1, 2027 under the California Constitution (scheduled, as of October 2026). Three details matter for UGC-style ads:

• Audio counts. The law's definition of an ad includes audio, so an AI-voiced narrator is covered.
• It names the wording. The label must be substantially similar to "this performance features a synthetic performer" or "no human performer is depicted."
• It targets prominent use. The bill defines "prominently" to cover a synthetic performer who demonstrates the product in the foreground, narrates the ad, or reacts on screen to its message.

Breaking it is treated as a violation of California's false advertising statute, section 17500. For example (hypothetical), a supplement brand running an AI presenter in nationwide ads could adopt California's wording as its single on-screen label for all cuts, then have a lawyer confirm it also meets New York's rule.

Voice and likeness cloning: get consent in writing

If an AI output looks or sounds like a real, identifiable person, the issue is consent.

• Tennessee, ELVIS Act (in effect since July 1, 2024, as of October 2026). It added voice to the state's protected personal rights. Knowingly using someone's voice in ads without prior consent, or publishing a voice or likeness you know is unauthorized, can lead to a civil action (Tennessee General Assembly).
• Hawaii, Act 247 (signed July 14, 2026 and in effect on approval, as of October 2026). Knowingly publishing a realistic AI imitation of an identifiable person in an ad without express, written consent is unlawful. The person can sue for damages of up to $25,000 per piece of content (HB 2137, Governor's office).
• California, Labor Code section 927 (in effect since January 1, 2025, as of October 2026). A contract clause that lets a company use a digital replica of someone's voice or likeness in place of work they would have done in person is unenforceable, with some exceptions, if it lacks a reasonably specific description of the intended uses and the person had no lawyer or union negotiating for them (California Legislature).
• Illinois, Digital Voice and Likeness Protection Act (in effect, as of October 2026). It has a similar rule for digital replica clauses in personal services agreements (Illinois General Assembly).

Federally, the NO FAKES Act of 2026 would create a right over voice and visual likeness with a notice-and-takedown process. It is a bill, not a law: the Senate Judiciary Committee reported it on June 24, 2026, and it sits on the Senate calendar as of October 2026 (Congress.gov).

How Meta, TikTok and YouTube label AI content

Platform rules decide whether your ad runs. All three platforms say they read C2PA Content Credentials metadata to apply AI labels automatically, so a file carrying it can be labeled even if you never self-disclose.

• Meta. Ads made or significantly edited with Meta's generative AI tools, or with third-party AI tools Meta detects, get an automatic "AI info" label. In some regions, including California and New York, advertisers can also self-disclose AI use so a label shows beside the ad label (Meta Business Help Center). For organic posts, Meta may require an AI label on photorealistic video or realistic-sounding audio, but not on images (Meta Help Center).
• TikTok. Creators must label AI-generated content that contains realistic images, audio or video (TikTok Support). In Ads Manager, advertisers switch on "This ad contains AI-generated content." Spark Ads follow the organic rules instead, so the creator's own post needs the label (TikTok Ads Help). Undisclosed AI content in ads will be rejected or restricted, and ads can't misuse a public figure's likeness for a false endorsement (TikTok ad policy).
• YouTube. Creators must select "Yes" under the "AI use" setting for realistic AI content, such as making a real person appear to say something they didn't. Beauty filters, AI-written scripts and cloning your own voice for a voiceover don't need disclosure (YouTube Help).

An AI label does not replace the FTC paid-partnership disclosure. You often need both.

Checklist for brands running AI-assisted UGC or creator campaigns

1. Ask about AI in the brief. Require creators and agencies to declare any AI used for faces, voices, footage or scripts. Jem's free campaign brief generator includes an FTC disclosure section; add your AI rules beside it.
2. Sort every asset into a bucket from the list above, so you know which rule applies.
3. Disclose the paid relationship on every sponsored post, in the video and the caption.
4. Label synthetic performers. New York already requires it, and California joins on January 1, 2027. For national ads, California's wording is the more specific standard.
5. Get written consent for any real person's voice or likeness, with a specific list of uses, platforms, duration and pay.
6. Never generate fake reviews or testimonials, and don't script an experience the creator didn't have. An AI presenter making a results claim needs the same proof a human would. Our health and wellness UGC guide shows how that proof standard works for health claims.
7. Use the platform AI toggles and re-check before boosting a creator post as a Spark Ad or through whitelisting: the paid disclosure and any AI label must be on the post itself.
8. Keep records. Save briefs, signed consents, AI tool declarations, drafts, approvals and final labeled files. The Endorsement Guides expect advertisers to guide, monitor and correct their endorsers (16 CFR 255.1(d)), and records show you did. A tracker with a rights log, like the one in our creative operations guide, makes this routine.

What creators should ask before a brand uses AI on their likeness

If a brand wants to clone your voice, build an AI avatar of you or extend your footage with AI, ask before you sign:

• What exactly will the AI version do? Get a specific list of uses. California and Illinois can void some vague replica clauses, but clear terms up front are easier.
• Where, and for how long? Platforms, paid or organic, and an end date.
• How am I paid for it? Replica use should be priced separately from the original shoot.
• Do I approve what it says? Ask for review rights on any script your AI version performs.
• What happens to the model and files after the deal? Ask for deletion or a clear end to the license.
• Can I rule out categories? For example, no political, health or competitor content.

California's rule specifically mentions legal counsel, so have a lawyer review any contract that licenses your digital replica.

Recent changes to AI regulations in the US for marketers

• December 11, 2025: Executive Order 14365 ordered a Justice Department task force to challenge state AI laws and called for federal preemption legislation (White House). Until a court or Congress acts on a specific law, plan to follow it.
• December 22, 2025: The FTC warned 10 companies about the Consumer Review Rule (FTC) and set aside its 2024 order against Rytr, an AI review-writing service (FTC). The rule itself still applies.
• June 9, 2026: New York's synthetic performer disclosure law took effect.
• June 24, 2026: The Senate Judiciary Committee reported the NO FAKES Act of 2026. It has not passed as of October 2026.
• July 1, 2026: The FTC published a proposed policy statement arguing that a state law like Colorado's AI Act is "impliedly preempted to the extent it conflicts with a federal regulatory scheme" (FTC). It is still a proposal as of October 2026.
• July 14, 2026: Hawaii's Act 247 on AI likeness imitations was signed.
• September 15, 2026: The FTC said its civil penalty amounts won't change in 2026 (Federal Register).
• September 16, 2026: California's SB 1050 was signed, effective January 1, 2027.

Run your next creator campaign with clean paperwork

The simplest way to lower AI risk is to know who made each asset, with what tools, under what terms. On Jem Social, you can find creators with the Influencer Finder, send a brief and contract, review drafts and request revisions before you approve, and pay through escrow (see what's included).

Put your AI disclosure and likeness terms in the brief. Starting with smaller creators? Our guide on how to find micro-influencers walks through the search.

Start your 7-day free trial of the Influencer Finder and book your first creators with disclosures and consents sorted from day one.