Do I have to tell US shoppers a product photo was made with AI?

Not under any general federal rule as of September 24, 2026. You must not mislead, and New York requires a disclosure when an ad shows an AI-generated person. Platforms such as Meta, TikTok and Google have their own labels; see the ad labeling guide.

Is "image for illustration purposes" enough to cover an AI image?

Often not. The FTC judges the net impression of the whole ad, and its Policy Statement says accurate text may not remedy a false headline. If the picture shows something the product does not do or include, change the picture.

Did the FTC drop its AI enforcement in 2025?

It reversed one case. On December 22, 2025 it reopened and set aside its 2024 order against Rytr, an AI review-writing tool, saying the complaint did not meet the law’s requirements. The fake reviews rule and the Endorsement Guides both still appear in the 2026 edition of the Code of Federal Regulations, and the FTC said it will keep acting against people who use AI to deceive.

Can an AI model show how my dress fits?

Yes, if the fit, drape and length match the real garment on a body of that size. An image that makes a garment look better than it fits can mislead. Base the image on real fit data and say which size is shown.

What are the FTC rules for AI product images?

Checked September 24, 20266 min read

Short answer

No federal rule says you must label an AI product photo in the US. The FTC Act bans ads that are likely to mislead, and the FTC judges the whole ad, pictures included. An AI image crosses the line when it shows the product looking, doing or including something the real one does not, or presents invented customers or reviews as real. New York has required a disclosure on ads with AI-generated people since June 2026.

The law that applies: deception under Section 5

Section 5 of the FTC Act declares "unfair or deceptive acts or practices in or affecting commerce" unlawful (15 U.S.C. § 45). The FTC’s 1983 Policy Statement on Deception sets three tests: a claim or practice likely to mislead, judged from the view of a consumer acting reasonably, about something material to the decision to buy.

Pictures are claims. The Policy Statement says the Commission may scrutinize the visual imagery of ads because it judges the impression an ad makes as a whole, and that accurate text may not fix a false headline. On that reasoning, a small caption is unlikely to fix a misleading image.

When the FTC announced its Operation AI Comply cases on September 25, 2024, it said there is no AI exemption from the laws already on the books. On December 22, 2025 it withdrew one of those orders, against the review-writing tool Rytr, saying the complaint did not meet the Act’s requirements, and repeated that it will act against people who use AI to deceive.

Mood picture or proof: the Colgate test

A 1965 Supreme Court case draws a useful line for AI product images. In FTC v. Colgate-Palmolive (April 5, 1965), a shaving cream ad claimed its product could shave sandpaper and showed it, using sand glued to plexiglass. The Court sided with the FTC: the undisclosed mock-up was a material deceptive practice, because the ad offered viewers visual proof of the claim.

The Court also said what it did not cover: an ad in which happy actors eat what looks like ice cream but is mashed potatoes. There the prop sets a mood. It becomes a problem when the ad invites viewers to see the product’s qualities for themselves. Ask of every AI image whether it sets a mood or offers proof of what the product is or does.

AI image	Mood or proof	Risk
Your real product, shown accurately, in an AI-made kitchen or beach scene	Mood	Low
A before-and-after skin result made with AI	Proof of results	High: results must be real and backed by evidence
A stain vanishing after one spray, generated by AI	Proof of performance	High
A garment on an AI model, showing how it fits	Proof of the product’s shape and fit	Medium: fit, drape and length must match the real item
A render that adds a strap, lid or accessory the box does not include	Shows what the buyer gets	High
An AI person holding the product and praising it	An endorsement	High unless the rules below are met

For category detail, see showing clothes on different body sizes and supplement photos.

AI people, reviews and endorsements

Fake reviews and testimonials. The FTC’s Consumer Reviews and Testimonials Rule (16 CFR Part 465), in force since October 21, 2024, bars a business from writing or selling a review or testimonial that misrepresents that the reviewer exists or used the product, and from spreading one it knew or should have known was fake. The FTC’s announcement named AI-generated fake reviews, and courts can impose civil penalties for knowing violations.
AI avatars. The FTC’s questions-and-answers page on that rule says there is no blanket ban on AI-generated avatars in marketing. A celebrity avatar used without permission breaks the rule if reasonable consumers would think the celebrity gave the testimonial.
Actual consumers. The Endorsement Guides, revised in July 2023, say an ad that presents people as actual consumers should show actual consumers, or clearly and conspicuously disclose that they are not (16 CFR 255.2(c)). An AI-made "customer" needs that disclosure.
Honest opinions. An endorsement must reflect the honest opinions or experience of the endorser, and an endorser shown using a product must have been a real user of it (16 CFR 255.1). A likeness can itself be an endorsement.

State laws on top of the FTC

New York added a specific AI rule. General Business Law § 396-b, signed on December 11, 2025 as Chapter 617 and in force 180 days later, on June 9, 2026, requires anyone who produces an ad to disclose conspicuously that it contains a synthetic performer, when they know it does. A synthetic performer is a digitally created human performer who is not recognizable as any real person. Penalties are $1,000 for a first violation and $5,000 for each one after. Audio-only ads, ads for expressive works such as films and games, and AI used only to translate a real performer’s speech are exempt.

Using a real person’s likeness is a separate question under state right-of-publicity laws; see AI images of real people. California’s AI Transparency Act, operative since August 2, 2026, puts its duties on large AI providers rather than on advertisers; see Content Credentials.

A pre-launch check for US ads

Put the AI image next to a photo of the real product. Shape, color, size, parts, label text and finish should match.
Check that everything shown comes in the box, or say clearly what does not.
If the image shows a result or a demonstration, confirm the product does that and that you can prove it.
If an AI-made person appears as a customer, disclose that they are not a real customer.
If the ad runs in New York and shows an AI-made person, add a conspicuous synthetic performer disclosure.
Never pair an AI face with a quote, star rating or review.
Make sure no face resembles a real person, and that no celebrity likeness appears without a signed agreement.
Keep the prompts, references and renders on file in case a platform or regulator asks.

Marketplaces add their own rules; see whether you can use AI images on Amazon.

Questions people also ask

Do I have to tell US shoppers a product photo was made with AI?: Not under any general federal rule as of September 24, 2026. You must not mislead, and New York requires a disclosure when an ad shows an AI-generated person. Platforms such as Meta, TikTok and Google have their own labels; see the ad labeling guide.
Is "image for illustration purposes" enough to cover an AI image?: Often not. The FTC judges the net impression of the whole ad, and its Policy Statement says accurate text may not remedy a false headline. If the picture shows something the product does not do or include, change the picture.
Did the FTC drop its AI enforcement in 2025?: It reversed one case. On December 22, 2025 it reopened and set aside its 2024 order against Rytr, an AI review-writing tool, saying the complaint did not meet the law’s requirements. The fake reviews rule and the Endorsement Guides both still appear in the 2026 edition of the Code of Federal Regulations, and the FTC said it will keep acting against people who use AI to deceive.
Can an AI model show how my dress fits?: Yes, if the fit, drape and length match the real garment on a body of that size. An image that makes a garment look better than it fits can mislead. Base the image on real fit data and say which size is shown.

Where Overs fits

Overs gives each photo only the reference pictures it needs, and on Pro and Team a reviewer checks each photo against those references before you see it. A person should still compare every photo with the real product before it runs, because the FTC judges what the published ad shows.

Free for 40 photos a month. The AI that makes the photos is billed separately, on your own key, with no markup from Overs.

Start free How it works

Sources

Checked on September 24, 2026. Prices, specs and rules change; follow the links for the current versions.