Amazon Now Makes You Label AI People in Your Images. Audit Your Catalog Before Q4.

If you swapped a photographer for an AI model somewhere in the last two years, Amazon just made that a labeled decision. On July 23, 2026, Amazon notified third-party sellers that any listing image or A+ content module showing a photorealistic AI-generated person now has to be tagged as synthetic. The trigger is New York’s synthetic performer disclosure law, S.8420-A, which Governor Hochul signed in December 2025 and which took effect June 9, 2026.

Here is the part that matters for your account: Amazon does not build a different catalog for shoppers in New York versus Texas. One catalog serves everyone. So a state law effectively becomes a marketplace-wide rule, and this one lands on your images, which are the single biggest lever you have on conversion. This is not a fee change and not a suspension threat on day one. It is a creative and compliance change that quietly touches the asset doing the most work on your detail page.

What actually changed

Amazon now expects you to disclose AI-generated people in your creative, not just avoid misrepresentation. The disclosure is technical, not a visible watermark you add yourself. You tag the image at the file level.

The mechanic works like this. Using an IPTC-compatible image editor, you add the keyword contains-synthetic-performer to the dc:subject field in the image’s XMP metadata before you upload it. Amazon reads that metadata and, where the rule applies, surfaces a shopper-facing indicator that the image contains a synthetic performer. The law defines a synthetic performer as a digital asset created or modified with generative AI that is meant to read as a human performance, without being a real, identifiable person. That definition is broad. An AI-generated lifestyle model holding your product is squarely inside it. A fully CGI product render with no human is not.

The disclosure obligation sits on the party putting the image into commerce, which is you. New York’s civil penalties run $1,000 for a first violation and $5,000 for each one after that. Amazon layering its own enforcement on top is the more immediate risk to your account, because Amazon does not wait for a court to act on a policy it has decided to enforce.

Why this bites conversion, and why Q4 makes it worse

The compliance math is small. The conversion math is not.

A shopper-facing “AI-generated” indicator on your hero or lifestyle image is a trust signal, and the early read on synthetic-content labels across platforms is that they tend to depress engagement rather than lift it. Buyers slow down when a badge tells them the person selling them the product is not real. You do not get to choose whether that label appears once your metadata says the image is synthetic. Amazon controls the surface.

Now put the calendar on it. Q4 is when your traffic and ad spend peak, so any conversion drag is multiplied by your highest-volume weeks. Discovering in mid-November that your best lifestyle image carries a trust-eroding badge, on the ASIN you are paying the most to promote, is the expensive version of this. The cheap version is finding it in August and reshooting.

There is also the reverse risk. If you used AI people and do not tag them, you are non-compliant and exposed to both New York’s penalty structure and Amazon’s enforcement. So the two failure modes are: label it and lose conversion, or hide it and carry policy and legal risk. Neither is a place you want to be blind to going into peak.

The audit to run this week

This is a catalog hygiene problem with a creative decision attached. Run it in order.

Inventory every image and A+ module that contains a person, across your active catalog. Pull your image assets and flag anything with a human figure. You are looking for AI-generated or AI-modified people specifically, so the fastest filter is by how the asset was produced: anything that came out of a generative tool or was retouched with generative fill on a person is in scope.

Separate true AI people from real photography and pure product renders. A CGI render of the product with no human does not trigger this. A real model shoot does not trigger this. Only synthetic humans do. Get this split right, because over-tagging real photography hands you a conversion penalty you did not owe.

For every confirmed AI-person image, make the call: tag or replace. On low-velocity ASINs, tag the metadata correctly and move on. On your top revenue and top ad-spend ASINs, seriously consider replacing the AI lifestyle shot with real photography before Q4, so no trust badge lands on the images you are about to pour spend behind.

Fix the tagging at the file level, not in the catalog manager. Add contains-synthetic-performer to the dc:subject XMP field with an IPTC-compatible editor, then re-upload. Build this into your normal creative workflow so new assets are tagged at creation, not audited after the fact.

Prioritize by money, not by count. A perfectly tagged catalog on products that barely sell is wasted effort. Start with your top-20% revenue ASINs and the SKUs carrying the heaviest Q4 ad budgets, because that is where a conversion swing actually moves your P&L.

The operator takeaway

The pattern here is the one worth internalizing. Amazon is increasingly the enforcement arm for regulation it did not write, and because it runs one catalog for the whole country, the strictest applicable state rule becomes your default. That will keep happening. The sellers who stay clean are the ones treating creative as tracked, owned infrastructure with a compliance layer, not a set-and-forget asset from launch.

The deeper signal is about AI-generated models themselves. They looked like a cheap shortcut to lifestyle imagery. Now they carry a disclosure cost and a likely conversion cost on your most important listings. For your hero and primary lifestyle images on revenue-driving ASINs, real photography just got more valuable, not less. Reserve the synthetic assets for places where a trust badge does not cost you a sale.

If you want a team that audits your full image catalog, sorts real photography from AI-generated people, tags the metadata correctly, and reshoots the hero images on your Q4 money-makers before peak, book a call with MarketplaceMax. We keep your creative compliant and your conversion intact heading into the quarter that pays for the year.