AI content disclosure rules in 2026: what advertisers must label

Do AI-generated ads have to be labeled in 2026?

Yes. In 2026, an advertisement that features AI-generated video, images, or a synthetic performer must be disclosed in most of the markets a US or EU brand cares about. Four overlapping rules now apply: the FTC Endorsement Guides (federal, in force today), New York's General Business Law Section 396-b (effective June 9, 2026), California's AI Transparency Act (SB 942, amended by AB 853, operative August 2, 2026), and the EU AI Act Article 50(2) (effective August 2, 2026). Each rule targets the same failure: passing off AI-made or paid content as organic, unlabeled, and human. None of them bans AI advertising. They require a clear, conspicuous label and, in some cases, machine-readable provenance data embedded in the file. For a brand, the practical takeaway is simple. If a real person did not make the ad, or if money changed hands, say so, plainly, where the audience will see it.

Which US laws require AI disclosure in 2026?

Three US layers require AI or paid-content disclosure in 2026: the FTC Endorsement Guides, New York's GBL Section 396-b, and California's AI Transparency Act. The FTC layer applies today and nationwide. Under the Endorsement Guides (16 CFR Part 255, revised 2023), every paid post is legally an advertisement. Whenever a material connection exists between an endorser and a seller that an ordinary viewer would not expect, that connection must be disclosed clearly and conspicuously (16 CFR Section 255.5(a)). The 2023 revision extended the definition of "endorser" to cover virtual and AI-generated influencers, so a synthetic creator promoting a product sits squarely inside the rule. The FTC's maximum civil penalty reached $53,088 per violation in 2025, though that ceiling applies to knowing violations rather than a first, non-egregious slip. New York and California then add synthetic-specific duties on top of the federal baseline.

New York's synthetic performer law (GBL 396-b)

New York's General Business Law Section 396-b, effective June 9, 2026, requires any advertisement that features a synthetic performer to conspicuously disclose that a synthetic performer is in the advertisement. A synthetic performer is a digitally created asset, generated by AI or a software algorithm, built to give the impression of a human performer who is not a recognizable, identifiable real person. The duty falls on whoever produces or creates the advertisement for commercial purposes, and only when that producer has actual knowledge a synthetic performer appears (platforms that merely distribute the ad are not on the hook). Penalties run $1,000 for a first violation and $5,000 for each subsequent violation. In plain terms, if a brand runs an ad with a fully AI-generated spokesperson in New York, the ad has to say a synthetic performer is in it.

California's AI Transparency Act

California's AI Transparency Act (SB 942, as amended by AB 853) becomes operative August 2, 2026, and binds the generative-AI provider rather than the individual advertiser. A covered provider, defined as a system with at least 1 million monthly California users, must attach a manifest disclosure (a visible, medium-appropriate label), embed a latent disclosure (metadata naming the provider, the system, and a timestamp), and offer a free AI-detection tool. Larger online platforms face provenance-surfacing duties starting January 1, 2027. Each violation carries a $5,000 penalty, and each day counts as a separate violation. For a brand, the important point is that a provider-level obligation means the tool used to make the content is expected to carry provenance forward, so the advertiser is not left bolting a label on after the fact.

What does the EU AI Act require?

The EU AI Act, Article 50(2), takes effect August 2, 2026, and requires providers of generative-AI systems to mark AI-generated or manipulated output in a machine-readable format, and to disclose deepfake and synthetic content to the people who see it. The marking has to be detectable by machines (embedded provenance), not just a caption that a re-upload can strip. Non-compliance can reach up to 15 million euros or 3% of worldwide annual turnover, whichever is higher. The rule binds the provider of the AI system and, for deepfakes and synthetic media, the party deploying it. Any brand advertising to EU audiences with AI-generated creative is inside this regime the moment the content reaches European viewers. Like the US laws, the EU AI Act does not prohibit AI advertising. It requires that AI content be marked as AI, in a way that survives the trip from the tool to the platform.

What do the platforms do automatically?

TikTok, Meta (Instagram and Facebook), and YouTube automatically apply an AI-made label when they detect machine-readable provenance in an uploaded file. These platforms read C2PA Content Credentials and IPTC provenance metadata, the same open standards the EU AI Act and California's law point toward. When a file carries that data, the platform stamps its own "AI-generated" or "made with AI" badge on the post without the creator doing anything. The catch is that a simple container flag is fragile. Many "is-AI" tags get stripped when a video is re-encoded or re-uploaded, so the label silently disappears. Durable disclosure has to be embedded as signed Content Credentials that ride with the pixels. For an advertiser, this means the disclosure that actually holds up is the one baked into the export, not the caption typed on the way out the door.

What should a brand do to stay clean in 2026?

To stay clean under the 2026 rules, a brand should do four things:

• Label AI-generated creative as AI.
• Disclose any paid or material connection with a clear "#ad" or "paid partner" line, placed where viewers see it first.
• Keep machine-readable provenance embedded in the exported file, not just in the caption.
• Never pass off a synthetic creator as a real, unaffiliated person.

The through-line across the FTC Guides, New York's GBL 396-b, California's AI Transparency Act, and the EU AI Act is identical: disclose the AI, disclose the money, and do it conspicuously. A brand that discloses by default carries no compliance debt, because every one of these laws is written to punish hiding it, not doing it. The cheapest posture is to treat disclosure as a built-in feature of the content pipeline rather than a manual step to remember on every post.

RealInfluencer is built for that posture. Every output is auto-labeled by design, and every export carries invisible, machine-readable C2PA Content Credentials, so the platforms that read Content Credentials apply their AI label automatically. Every creator on RealInfluencer is fully synthetic, with no real people involved and all 18 or older, which keeps the content clear of the identifiable-real-person line the synthetic performer laws draw. Disclosure is treated as a feature, not a cost, so the labeling the FTC, New York, California, and the EU all now expect is already handled the moment content leaves the system.

This guide is general information, not legal advice, and a brand with specific exposure should confirm its obligations with counsel.

Start creating at realinfluencer.ai →