Amazon’s New AI Image Rule: What Sellers Need to Tag

Amazon sellers using artificial intelligence to create listing visuals now have a new compliance requirement to consider. Under the updated Amazon AI image policy, sellers must identify product images, videos, and A+ Content that feature photorealistic people generated entirely by AI.

 

The change does not prohibit sellers from using AI-generated product images. Instead, it introduces a disclosure process designed to help shoppers recognize when a person shown in commercial content is not real.

 

Amazon’s decision follows the implementation of a new law in New York requiring advertisers to disclose the use of AI-generated “synthetic performers.” For sellers, this means that synthetic performer disclosure must now become part of the content creation and listing upload process.

 

 

What Does Amazon’s New Policy Require?

Amazon has instructed third-party sellers to tag images and videos containing photorealistic AI-generated people before uploading them to product listings or A+ Content.

The requirement applies across Amazon’s worldwide stores, not only to listings specifically created for customers in New York.

 

To comply, sellers must use an IPTC-compatible metadata editor and add the following keyword to the dc:subject XMP metadata field of the applicable image or video file:

• contains-synthetic-performer
• Once Amazon detects this metadata, it may add an indicator to the product detail page informing customers that the visual content includes AI-generated people.

 

Amazon has not yet provided detailed criteria explaining exactly when or where the customer-facing indicator will appear. However, the company has confirmed that the metadata must be included before affected files are uploaded to listings and A+ Content.

 

Which Images and Videos Must Be Tagged?

The new Amazon AI image policy applies when an image or video includes a person who meets both of the following conditions:

• The person was generated using AI or another digital technology.
• The person appears photorealistic and could reasonably be mistaken for a real human.

For example, a seller using an AI tool to create a realistic-looking model wearing a jacket, holding a kitchen appliance, using fitness equipment, or demonstrating a beauty product would generally need to tag the file.

 

The requirement may also apply when an AI-generated person appears in lifestyle photography, comparison graphics, product demonstrations, brand storytelling videos, or enhanced A+ Content.

 

The key question is not whether AI was used somewhere in the production process. Sellers should determine whether the final visual includes a realistic-looking person who does not actually exist.

 

Which Types of Content Are Exempt?

Amazon has clarified that sellers do not need to add the metadata when the image or video:

• Features only real people, even when AI tools were used to edit or enhance them
• Includes characters from movies, television programs, streaming content, documentaries, or video games
• Does not contain any people
• Includes people who are clearly illustrated, animated, or otherwise not photorealistic

This distinction is important. Using AI to remove a background, improve lighting, adjust colors, retouch a real model, or create a product-only setting does not automatically trigger the disclosure requirement.

 

Likewise, an AI-generated illustration or cartoon character generally falls outside the policy when it could not reasonably be mistaken for a real person.

The policy is specifically focused on realistic synthetic people rather than every form of AI-assisted creative work.

 

Why Is Amazon Introducing This Requirement?

The policy change is closely connected to New York’s synthetic performer disclosure law, which took effect on June 9, 2026.

 

The law requires advertisements to clearly disclose when they use AI-generated synthetic performers in place of human actors. New York defines these performers as digitally created media that appear to be real people.

 

According to the New York State government, AI-generated performers and manipulated media may make it more difficult for consumers to distinguish real people from synthetic content when no disclosure is provided.

 

Governor Kathy Hochul described the legislation as the first state law of its kind in the United States. The measure is intended to improve advertising transparency, protect consumers, and address concerns about AI-generated performers replacing human creative workers.

 

Advertisements that fail to provide a conspicuous disclosure may face a $1,000 penalty for a first violation and penalties of $5,000 for subsequent violations. Certain expressive works, including movies, television programs, streaming content, and video games, are exempt when synthetic characters are used in connection with the underlying work.

 

What Should Amazon Sellers Do Now?

Sellers that use generative AI for product photography or marketing content should review their creative workflow rather than waiting until a listing is rejected or flagged.

1. Audit Existing Creative Assets

Review listing images, A+ Content, product videos, Store content, and other visual materials for photorealistic AI-generated people.

Amazon has indicated that compliance is mandatory for new uploads. Although sellers have not been instructed to immediately replace every previously uploaded file, updating older affected content is a sensible precaution, especially when a listing is already being edited.

 

2. Confirm How Each Person Was Created

Maintain clear records showing whether a person in an image is:

• A real model
• A real model edited with AI tools
• A completely AI-generated person
• A non-photorealistic illustrated character

This information can help marketing teams, agencies, designers, and account managers apply the policy consistently.

 

3. Add the Required Metadata Before Uploading

Do not rely on the image’s filename, alt text, listing copy, or internal file notes. Amazon specifically requires the keyword to be placed in the appropriate XMP metadata field using an IPTC-compatible editor.

The required keyword is:

contains-synthetic-performer

After adding the metadata, sellers should verify that it remains embedded in the final exported file. Some image-compression platforms, editing programs, and content management tools may remove metadata during processing.

 

4. Review Content Supplied by Agencies and Freelancers

Many sellers outsource product photography, A+ Content, advertising graphics, and video production. Contracts and creative briefs should now require outside providers to disclose whether synthetic people were used.

Sellers should not assume that an agency has embedded the correct metadata simply because it used an AI image generator.

 

5. Update Internal Content Guidelines

Brands producing large volumes of AI-generated product images should add a compliance checkpoint before files are approved for upload.

A simple internal review can ask:

• Does the asset include a person?
• Is that person entirely AI-generated?
• Does the person appear photorealistic?
• Has the required metadata been added?
• Has the final exported file been checked?

This process can help prevent incorrectly tagged files as well as missed disclosures.

 

Amazon Continues Expanding Its Own AI Tools

The new labeling requirement comes as Amazon continues to invest heavily in artificial intelligence across its retail and advertising businesses.

 

Amazon Ads already offers generative AI solutions that can help advertisers create images, videos, audio, and campaign materials. Its Creative Agent and Creative Studio tools are designed to streamline tasks ranging from product research and copywriting to storyboarding and full video production.

 

Third-party sellers are also increasingly using AI to write product descriptions, produce lifestyle visuals, generate advertising concepts, and adapt content for different markets.

 

Amazon’s latest rule therefore should not be viewed as a rejection of AI-generated content. It signals that platforms are beginning to separate the use of AI from the obligation to disclose realistic synthetic people.

Sellers may continue benefiting from faster and more affordable content production, but they will need better oversight of how that content is created and presented.

 

AI Disclosure Is Becoming a Broader Platform Standard

The United States does not yet have one uniform federal framework governing every type of AI-generated advertisement. Instead, states are developing their own requirements, creating a more fragmented compliance environment for brands selling nationwide.

 

Other major online platforms have also introduced or proposed labels for AI-generated or digitally altered content. These developments suggest that transparency requirements are likely to become more common across e-commerce, advertising, and social media.

 

For international sellers, this trend presents an additional challenge. A product listing may be created in one country, managed by a team in another country, and displayed to customers across multiple U.S. states. Content policies should therefore be designed around the strictest relevant requirements rather than the seller’s physical location alone.

 

Final Takeaway for Amazon Sellers

The updated Amazon AI image policy does not ban generative AI or prevent sellers from using AI to improve their listings. Its main purpose is to ensure that shoppers are informed when product content includes a photorealistic person who was created digitally rather than photographed in real life.

 

Sellers should review their AI-generated product images, identify realistic synthetic people, and embed the required metadata before uploading new listing or A+ Content files.

 

By building synthetic performer disclosure into the creative approval process, Amazon sellers can continue using AI efficiently while reducing compliance risks and maintaining transparency with customers.

 

As AI regulations continue to develop, clear records, careful content reviews, and consistent disclosure practices will become increasingly important parts of Amazon account management.