Amazon just told millions of third-party sellers they have to label any product image that features an AI-generated person. The rule is not a suggestion. It is a policy change that went into effect immediately after a new New York law forced the issue, and it is going to reshape how a lot of e-commerce listings look from here on out. AI-generated images have been creeping into product pages for over a year, and Amazon is finally drawing a line.
What actually changed
Starting in late July 2026, Amazon requires all marketplace sellers to tag AI-generated images and videos containing AI-generated people with specific metadata keywords before uploading. The rule covers product photos, A+ content (the enhanced listings with videos and extra graphics), and any promotional media on a listing page.
When a listing contains tagged AI-generated images, Amazon will add an indicator on the page that shoppers can see. The label tells consumers the person in the image is not real. Amazon has not specified exactly what the indicator looks like or when it will appear, but the company confirmed it will be visible “where applicable.”
The rule does not apply to everything. If your product photo shows a real person, even one that has been slightly retouched or enhanced with AI, you do not need to label it. Images of fictional characters from TV shows, movies, or video games are also exempt. The requirement specifically targets photorealistic AI-generated images of humans that could be mistaken for real people.
Why this is happening now
New York passed a first-in-the-nation law requiring companies to disclose “synthetic performers” in advertising. The law took effect in June 2026, and it applies to any “digitally-created media that appear as a real person” used in ads. Governor Kathy Hochul described the problem clearly: without disclosure, AI-generated content “can undermine one’s ability to accurately distill fact from fiction.”
Amazon is not the only platform scrambling to comply with rules around AI-generated images. Meta, TikTok, Pinterest, and YouTube have all added AI content labels. California also started requiring large AI providers to embed watermarks in generated content this year. The regulatory pressure is building fast, and Amazon’s move is less about corporate responsibility and more about legal exposure.
The seller problem
Here is the thing a lot of coverage misses. Amazon sellers have been using AI-generated images for product photography for over a year now, often through Amazon’s own tools. The company has been actively encouraging sellers to use AI to create listing content, product photography included. Now the same sellers are being told to label the AI-generated images they were incentivized to create.
For small sellers running dozens or hundreds of listings, going back and tagging every AI-generated image is a non-trivial task. And if you used a tool that does not clearly mark which images are AI-generated versus which are real photos, figuring out which ones need labels could be a mess. The metadata tagging system Amazon set up relies on self-reporting. Uploaders are expected to add the required keywords themselves before submitting.
What this means for e-commerce sellers using AI
If you are selling on Amazon and using AI tools to create AI-generated images, you need to change your workflow. Start tagging content at the point of creation, not after the fact. Keep a spreadsheet or folder system that tracks which images are AI-generated and which feature real models or no people at all.
The disclosure requirement is not going away. If anything, it is going to expand. Other states are drafting similar laws, and federal legislation has been proposed. Getting your tagging system in order now saves you from a scramble when the next rule drops.
Does labeling hurt conversions?
That is the question every seller is asking. The honest answer: we do not know yet. Some data from social media platforms suggests that AI content labels do not significantly change engagement rates when the content is high-quality. People care more about whether the product looks good than whether the person holding it is real.
But product images are different from social posts. Shoppers on Amazon are in buying mode, and trust matters more here. A label that says “this person does not exist” could subtly shift how someone evaluates a product, especially for items where the human element matters, like clothing, cosmetics, or fitness gear.
The safest approach is to use AI for background generation, product placement, and scene composition while keeping real people (even stock photo models) for the human elements of your listings. That way you get the efficiency of AI-generated images without needing the disclosure label at all.
The bigger picture
Amazon’s new rule is not really about AI-generated images specifically. It is about the growing gap between what AI can create and what consumers expect to be real. When you see a product listing with a smiling person holding a supplement bottle, you assume that person exists. AI erases that assumption, and regulators are stepping in to put it back.
For sellers, the playbook is simple. Stay honest, tag what needs tagging, and do not cut corners on disclosure. The platforms are going to keep tightening these rules, and the sellers who adapt early will be the ones who are not scrambling when the next policy update lands in their inbox.
