AI Image Disclosure Laws in 2026: A Brand's Compliance Guide
Prateek Gupte
Yes, in 2026 most brands using AI-generated imagery in advertising have to disclose it. New York’s synthetic performer law is already in effect. The EU AI Act’s labeling rules land August 2. And the FTC can treat undisclosed AI in an ad as a deceptive practice anywhere in the US, with penalties north of $53,000 per violation. If an AI-generated person appears in your ad, the safe default in 2026 is to label it.
This guide maps the four laws that actually matter for ecommerce and fashion brands, shows you where disclosure applies (ads, yes; product pages, it depends), and walks through how to label AI images without making your creative look like a warning sticker.
Key Takeaways
New York is live. N.Y. Gen. Bus. Law § 396-b requires conspicuous disclosure of AI synthetic performers in ads reaching NY consumers, in effect since June 9, 2026. Penalty: $1,000 first violation, $5,000 after.
The EU AI Act lands August 2, 2026. Article 50 requires AI-generated images to be machine-readable and labeled when published. A pending transition may push some providers to December 2.
The FTC is the real teeth in the US. Undisclosed AI in ads can be a deceptive practice under Section 5, with a maximum penalty of $53,088 per violation as of January 2026.
A generated model is not a digital twin. NY’s law covers synthetic performers who aren’t identifiable real people. Using a real person’s AI likeness triggers separate right-of-publicity rules.
Disclosure builds trust. AI labels have been shown to lift consumer trust, so the brands treating this as a values statement, not a compliance chore, are winning.
When Maya, who runs marketing for a mid-market apparel brand, swapped her fall campaign to AI-generated models in May 2026, her first instinct was to keep quiet about it. Her agency disagreed. They added a single line, “Models in this campaign were created with AI,” to the paid social creative running in New York. The campaign performed fine. What did not happen is the part that matters: no consumer complaint, no scramble when the law took effect three weeks later, no awkward call with legal. The disclosure cost her one line of copy. Skipping it could have cost $5,000 a violation.
Do You Actually Have to Disclose AI-Generated Images?
In most advertising contexts in 2026, yes. The clearest trigger is simple: if an AI-generated human appears in a commercial ad, you almost certainly need to disclose it. That covers AI fashion models, synthetic spokespeople, and AI-generated influencers in paid placements.
The grey areas are everything else. AI used only to remove a background, correct color, or relight a real photo generally sits below the disclosure line, because the image still depicts a real product and real people. A fully AI-generated scene with an AI model sits clearly above it. The practical rule: disclose when AI created or substantially changed what the viewer sees, especially a person.
Three questions sort almost any image:
Is it an advertisement? Paid social, display, and video ads are squarely in scope. Organic product pages are a softer case.
Does it show an AI-generated person? A generated human is the highest-risk element across every law on this list.
Does it reach a regulated market? New York, the EU, and California each pull in any brand whose content reaches their consumers, wherever you are based.
If you answer yes to all three, disclose. For a deeper read on the rights and ethics underneath all of this, our complete legal and ethical guide to AI fashion imagery goes section by section.
The Four AI Image Disclosure Laws That Matter in 2026
Four regimes shape what brands must do:
NY § 396-b — AI synthetic performers in ads reaching NY consumers. In effect (June 9, 2026). Penalty $1,000 / $5,000.
EU AI Act, Article 50 — AI-generated images published in the EU. Applies August 2, 2026. Fines up to 7% of turnover in severe cases.
California SB 942 — large generative AI tools (over 1M users). Operative August 2, 2026.
FTC Act, Section 5 — deceptive AI in US advertising. Always on. Up to $53,088 per violation.
New York: the Synthetic Performer Disclosure Law
Codified as N.Y. General Business Law § 396-b (bills S8420A / A8887B), in effect since June 9, 2026, it requires anyone producing a commercial ad to conspicuously disclose an AI-generated synthetic performer: a digital figure created by generative AI that is not recognizable as any identifiable real person. Penalty: $1,000 first, $5,000 after. It reaches any brand whose ads touch New York. We cover the mechanics in our explainer on New York’s AI disclosure law.
The EU AI Act: Article 50 transparency
From August 2, 2026, providers of generative AI must mark outputs as machine-readable AI, and deployers (including brands that publish the content) must label AI-generated or manipulated images. A pending Digital Omnibus agreement may give some tools a transition window to December 2, 2026. The full text lives in Article 50 of the EU AI Act.
California: the AI Transparency Act (SB 942)
Operative August 2, 2026 (moved from January via AB 853). It targets the tools: generative AI providers with over a million monthly users must offer a visible disclosure plus a hidden, latent one (metadata or watermark) and a free public detection tool. It determines what provenance data rides along with the images you generate.
The FTC: the federal backstop with real teeth
Under Section 5 of the FTC Act and the updated Endorsement Guides, undisclosed AI in advertising can be a deceptive practice. In January 2026 the FTC stood up a dedicated AI enforcement unit and raised the maximum penalty to $53,088 per violation, nationwide. Its AI guidance hub is the canonical source.
Want output built to be disclosed cleanly? See how Caimera handles rights and provenance.
Synthetic Performers vs. Digital Twins: Two Different Rulebooks
A synthetic performer is a generated person who is not a real, identifiable individual, an AI fashion model that exists only as pixels. New York’s § 396-b and the EU AI Act’s labeling rules are built for exactly this: disclose that the figure is AI.
A digital twin is an AI recreation of a specific real person. The moment a real, identifiable likeness is involved, you leave disclosure-law territory and enter right of publicity and digital-replica law, which turns on consent and licensing, not labeling. You generally need the person’s permission before you generate or run their likeness.
The cleanest path is to use generated models for scale and reserve digital twins for licensed talent. Caimera supports both: AI fashion models you can pick from a curated model library, plus Enterprise digital twins with the rights handled. Never generate a real person’s likeness without consent.
Where Disclosure Applies: Ads, Product Pages, and Social
Paid social and display ads are the clearest case. Instagram, Meta, TikTok, and Pinterest placements with an AI-generated person are in scope under New York’s law and will need labeling under the EU AI Act.
Organic product pages are the genuine grey area. New York’s law targets advertisements, and a standard PDP may not qualify, but a promotional page can be read either way. Get a lawyer’s read on your specific pages.
You are responsible for your agency’s work. If a partner makes an undisclosed AI choice on your campaign, the brand carries the liability. Update vendor contracts to require disclosure.
How to Disclose AI Images Without Killing Your Creative
Across all four regimes, a handful of mechanisms satisfy the requirements, and most are invisible to the casual viewer:
A visible label or caption. A short line like “Created with AI” near the creative, placed where people will see it.
Alt text and product-description tags. An “AI-generated image” note that doubles as accessibility and SEO value.
Embedded metadata. A field such as ai_generated=true, the backbone of the EU’s machine-readable requirement.
C2PA / Content Credentials. Cryptographic provenance that travels with the file, the default way to prove origin at scale.
A persistent watermark. A small AI mark in a corner, useful for video and high-visibility placements.
For most brands, the winning combination is a visible caption on ads plus embedded metadata on everything. The caption satisfies New York and the FTC; the metadata satisfies the EU and gives you an audit trail.
Want disclosure-ready assets out of the box? Enterprise plans include legal indemnification for AI content, the line that gets legal and procurement to yes.
Platform Rules Often Bite Before the Law Does
The platforms where your creative runs have their own AI-disclosure policies, and they often enforce faster than any regulator:
Meta (Instagram and Facebook) applies AI info labels and asks advertisers to disclose AI-generated imagery in certain contexts.
TikTok auto-labels AI-generated content and requires creators to flag realistic AI media.
Pinterest labels AI-generated Pins and uses metadata signals to detect them.
Amazon has tightened its stance on AI imagery in listings, especially main images.
Shopify permits AI-generated product imagery and leans on disclosure and accuracy rather than prohibition.
Even if a statute would let you skip a label, the platform may not. Build disclosure into your workflow once and you satisfy both.
Turn Compliance Into a Trust Advantage
Disclosure is not just risk management. In a Yahoo and Publicis Media study, adding an AI disclosure label was associated with a lift in overall consumer trust, not a drop. A majority of shoppers say AI makes it hard to know what is real, and most already assume brands use AI in their ads. Silence reads as hiding. A clear, on-brand label reads as confidence.
The brands winning here treat the label as a values statement. That is the same instinct behind keeping AI output on-brand instead of shipping slop. The cautionary flip side, what happens when a platform’s trust collapses, is worth reading in our Higgsfield case study.
Your 2026 AI Disclosure Checklist
Audit your live creative in New York, the EU, and California for AI-generated people.
Set a single disclosure standard: visible caption on ads, embedded metadata everywhere.
Separate synthetic performers from digital twins, and confirm consent for any real likeness.
Update agency and vendor contracts to require AI disclosure on work they deliver.
Draft your label language and get a lawyer to bless it. Test a couple of phrasings.
Choose tools that produce provenance metadata so the EU and FTC trails are automatic.
Frequently Asked Questions
Do I have to disclose AI-generated images on my website?
For advertising, yes in New York and across the EU once Article 50 applies. For organic product pages it is a grey area: New York’s law targets advertisements, and a standard PDP may not qualify, but a promotional page can be read either way. The safe default is to disclose any AI-generated person and get legal advice on your specific pages.
What happens if I don’t disclose AI in my ads?
In New York, $1,000 for a first violation and $5,000 for each one after. Federally, the FTC can treat undisclosed AI as deceptive with penalties up to $53,088 per violation. In the EU, AI Act breaches can reach a percentage of global turnover in severe cases.
Does AI background removal or color correction count?
Generally no. Minor edits to a real photo sit below the disclosure line because the image still shows a real product and real people. Disclosure kicks in when AI creates or substantially changes what the viewer sees, especially a person.
Is an AI fashion model the same as a deepfake?
No. An AI fashion model is a synthetic performer, a generated person who is not a real individual. A deepfake recreates a specific real person’s likeness, which triggers right-of-publicity and digital-replica law and usually requires that person’s consent.
What should an AI disclosure actually say?
None of the laws prescribe exact wording. Short, specific, and visible works best: “Created with AI” or “AI-generated model.” Place it where viewers will see it and apply it consistently. Vague or buried labels can fail the conspicuous standard.
When does the EU AI Act labeling rule start?
August 2, 2026 for the core transparency obligations. A pending Digital Omnibus agreement may give some providers a transition window to December 2, 2026, so confirm the status for your specific tools.
The Bottom Line
AI image disclosure laws arrived faster than most brands planned for, and 2026 is the year they get real. New York is already live, the EU AI Act lands in August, California follows, and the FTC has been watching all along.
The fix is not complicated. Set one disclosure standard, separate generated models from licensed digital twins, fix your vendor contracts, and pick tools that carry provenance metadata. And remember the upside: a clear AI label earns trust rather than losing it. The label is not a warning. In the hands of a confident brand, it is a statement.
This guide is general information, not legal advice. AI disclosure laws are changing quickly, so confirm the current rules with qualified counsel before you publish.
Ready to produce on-brand AI imagery that is built to be disclosed? Start free with 50 credits, compare plans, or talk to our team about Enterprise indemnification.
