As confirmed in Amazon’s Seller Central announcement, any product image or A+ Content video containing a photorealistic AI-generated person must be tagged with a specific metadata keyword before upload to any Amazon store worldwide, and Amazon moderators have confirmed that the requirement has been mandatory since June 6, 2026, which means sellers using Amazon AI generated image assets with photorealistic people in their listings are potentially non-compliant right now even if they never received a notification about the change.
This Amazon AI generated image metadata requirement is legally grounded in New York General Business Law Section 396-b, which took effect June 9, 2026 and mandates disclosure when advertising uses a synthetic performer in place of a real person. Amazon implemented the policy globally rather than limiting it to New York-facing listings. The practical compliance step for any Amazon AI generated image is specific and technical: sellers must add the keyword contains-synthetic-performer to the dc:subject XMP metadata field of the image or video file using an IPTC-compatible editor before uploading to Seller Central. This post covers exactly which Amazon AI generated image assets trigger the requirement, what the tagging process involves, which tools handle the metadata correctly, what Amazon does with the tag once it receives it, and what the enforcement consequence is for qualifying images that lack the tag.
Why sellers need to check their existing catalog now
Amazon moderators recommend retroactively updating existing qualifying content, meaning images already live in listings before the announcement are not exempt. Any seller who has used AI lifestyle imagery featuring people should audit their catalog before Q4 traffic peaks.
Amazon AI image metadata requirement at a glance, confirmed in Amazon’s Seller Central announcement:
- What triggers it: Any product image or A+ Content video containing a photorealistic AI-generated person
- Exact keyword:
contains-synthetic-performer - Where it goes: The
dc:subjectXMP metadata field - How to add it: Using an IPTC-compatible metadata editor before uploading to Amazon
- What Amazon does: Adds a customer-facing disclosure indicator on the detail page
- Applies to: All Amazon worldwide stores, product listings, and A+ Content
- Mandatory since: June 6, 2026 per Amazon moderator confirmation; announcement published July 22, 2026
- Retroactive: Amazon moderators recommend updating existing qualifying content
Table of Contents
- Which Images Trigger the Requirement
- What Is Specifically Exempt
- The New York Law That Drove Amazon’s Global Policy
- The Tagging Process: Step by Step
- Which Tools Handle XMP Metadata Correctly
- A+ Creative Studio and Automatic Tagging
- What Happens to Non-Compliant Images
- How to Audit Your Existing Catalog
- Frequently Asked Questions
- How DAM Law Firm Can Help
Which Images Trigger the Requirement
Product images and A+ Content videos that contain a photorealistic person AI generated entirely and not based on a real, identifiable person trigger the requirement. For any given image, ask whether the depicted person is a photorealistic synthetic human: created by AI rather than photographed from a real person.
The definition of a synthetic performer under the New York law
The New York legislation defines a synthetic performer as digitally created media that gives the impression of a human performance when the figure is not recognizable as an identifiable natural performer. For Amazon’s purposes, this translates to a photorealistic AI-generated person in a product image who looks like a real human but was created entirely by AI without being derived from any real, identifiable person’s likeness. A fully AI-generated model wearing a shirt in a lifestyle shot triggers the requirement. An AI-generated landscape background with no people does not.
Practical examples that trigger the requirement
Specifically, the most common situations that trigger the requirement in a seller’s catalog are AI-generated lifestyle images where a synthetic person is shown using or wearing the product, A+ Content modules featuring AI-generated people in lifestyle contexts, and product videos where a synthetic performer demonstrates product use. Sellers who have used AI image generation tools to create product imagery at scale should assume that any image featuring a human figure warrants review against the requirement. AI generation is faster and cheaper than photoshoots, but the compliance step cannot be skipped.
What Is Specifically Exempt
Amazon’s announcement is explicit about four exemption categories. Understanding what is exempt is as important as understanding what triggers the requirement, because misidentifying an exempt image as requiring a tag wastes time, and misidentifying a requiring image as exempt creates non-compliance risk.
Real people, even AI-altered
Amazon says media containing only real people does not require the metadata even if AI tools were used to alter them. The requirement targets an entirely AI-generated, realistic-looking person, not every use of AI in a product image. For example, a product photo of a real model where AI was used for background removal, skin retouching, or lighting adjustment does not trigger the requirement. The question is whether the person depicted is real, not whether AI was involved in producing the image.
Movie, TV, and video game characters
Amazon’s announcement specifically exempts characters from expressive works including movies, television programs, and video games. A product image featuring a recognizable character from a licensed franchise does not trigger the synthetic performer requirement even if that character is depicted using photorealistic rendering.
Non-photorealistic figures
Illustrated figures, cartoon characters, stylized avatars, and any human representation that is clearly non-photorealistic are exempt. The requirement targets images where the AI-generated person could plausibly be mistaken for a real human in a photograph. Abstract or clearly stylized human representations do not meet that threshold.
Images with no people
Product images that contain no human figures of any kind are fully exempt. A pure product shot on a white background, an AI-generated lifestyle environment with only the product and no people, and any image where human figures are absent do not require the metadata tag regardless of how extensively AI was used to create them.
The New York Law That Drove Amazon’s Global Policy
The reason this showed up now is legislation, not a change of heart from Amazon about AI art. New York passed an AI advertising disclosure law, New York General Business Law Section 396-b, that took effect on June 9, 2026. The law requires companies to disclose when advertising uses a synthetic performer instead of a real person. Amazon implemented the policy across all its worldwide stores rather than limiting it to New York-facing content, making it a global compliance requirement for all sellers in all marketplaces.
Why Amazon went global rather than New York-only
Notably, Amazon’s decision to implement the requirement globally reflects the direction of disclosure regulation broadly: New York’s law is one of the first, but similar requirements are under consideration or in development in other jurisdictions. Implementing globally rather than attempting to serve different metadata requirements to different regional storefronts avoids a fragmented compliance system that would be operationally complex and difficult to audit. For sellers, the practical implication is that complying with the requirement for all qualifying images regardless of which marketplace they appear in is both simpler and more durable than attempting to identify which images go to which stores.
The Tagging Process: Step by Step
Compliance requires adding metadata to the image file before it is uploaded to Amazon. The metadata must be in the file itself, not added through Seller Central after upload. Amazon reads the XMP metadata from the file during the upload process and uses it to determine whether to add a customer-facing disclosure to the listing.
Step 1: Identify qualifying images
Review every product image and A+ Content image or video in the catalog. For each image containing a person, determine whether the person is real (photographed) or AI-generated (synthetic). Any image containing a photorealistic AI-generated person is a qualifying image. Document the file names and locations of all qualifying images.
Step 2: Open the file in an IPTC-compatible metadata editor
The metadata must be added to the dc:subject field in the file’s XMP metadata using an IPTC-compatible editor. For qualifying images or videos, Amazon requires sellers to use an IPTC-compatible metadata editor and add the exact keyword: contains-synthetic-performer to the dc:subject (XMP) field of the media file. Sellers must add the tag before uploading the file to Amazon. Several tools handle this correctly, as described in the next section.
Step 3: Add the exact keyword
In the IPTC/XMP keywords or subject field, add the exact text: contains-synthetic-performer. The keyword must be exact: no variations, no additional text, no spaces added. Amazon’s system reads for this specific string and will not recognize a paraphrased version. Existing keywords in the field can remain; the synthetic performer keyword is added to the existing list, not substituted for it.
Step 4: Verify the tag was written correctly
On the command line, run exiftool -G1 yourfile.jpg and look for a line that reads [XMP-dc] Subject: contains-synthetic-performer. This verification step confirms the keyword was written to the correct XMP field before the file is uploaded. An image that appears to have been tagged but where the keyword was written to a different metadata field (IPTC keywords rather than XMP dc:subject, for example) will not satisfy the requirement.
Step 5: Upload the tagged file to Seller Central
Upload the tagged file through the normal image upload process in Seller Central. Amazon reads the metadata during the upload process. Once the metadata is detected, Amazon adds a disclosure indicator to the product detail page informing customers that the image includes AI-generated people, where applicable. The seller does not control the disclosure indicator’s appearance; Amazon generates it automatically.
Which Tools Handle XMP Metadata Correctly
Several widely used tools can add the required keyword to the correct XMP field. The choice depends on the seller’s existing workflow, technical comfort level, and the volume of images requiring tagging.
ExifTool (command line)
Specifically, ExifTool is a free, open-source metadata editor that handles XMP fields with precision. The command to add the synthetic performer keyword to a single file is: exiftool -XMP-dc:Subject+="contains-synthetic-performer" yourfile.jpg. For bulk processing of multiple files in a directory, ExifTool supports batch operations that apply the same metadata change across all files matching a pattern. ExifTool is the most reliable option for verifying that the keyword was written to the correct field, because it provides granular field-level output that confirms both the field and the value.
Adobe Bridge
Adobe Bridge, included with Creative Cloud subscriptions, provides a graphical interface for editing XMP metadata. In Bridge, open the Metadata panel, locate the IPTC Core section, and add contains-synthetic-performer to the Subject field. This writes to the XMP dc:subject field, the correct destination for Amazon’s requirement. Bridge also supports batch metadata application across multiple selected files, making it practical for sellers who need to tag a large number of images without command-line tools.
Adobe Lightroom
Lightroom’s keyword and metadata features can write to the XMP subject field. Adding contains-synthetic-performer as a keyword in Lightroom and then exporting the image with the metadata embedded will produce a file with the correct tag. Sellers using Lightroom for their product image workflow can integrate the tagging step into their existing export process. Verify with ExifTool after export to confirm the keyword was written to the correct XMP field rather than a Lightroom-internal field.
IPTC Photo Metadata Editor (browser-based)
The IPTC organization provides a free browser-based metadata editor at iptc.org that allows users to add and edit XMP metadata fields without installing software. This is the most accessible option for sellers who process images infrequently and do not have ExifTool or Adobe tools available. The editor exports a tagged version of the image file that can then be uploaded to Amazon.
A+ Creative Studio and Automatic Tagging
Amazon moderator guidance says qualifying content created through Amazon A+ generative AI features such as Creative Studio Amazon tags automatically. Sellers who generate AI-enhanced images entirely within Amazon’s Creative Studio tools do not need to manually add the metadata tag, because Amazon’s system applies it automatically to content it generates. This covers only content generated within Amazon’s own tools. Content generated externally using third-party AI image generation platforms (Midjourney, DALL-E, Stable Diffusion, Adobe Firefly, and similar tools) require manual tagging before upload regardless of how it is processed after generation.
What Happens to Non-Compliant Images
Importantly, Amazon has not published a specific enforcement timeline for images that are non-compliant with the synthetic performer metadata requirement. However, the enforcement mechanism that applies to any image that violates Amazon’s image requirements is listing suppression. The product Amazon removes the product from search results until a compliant image is supplied.
How non-compliance is identified
Amazon’s image review systems can identify Amazon AI generated image content through the same detection technology that the requirement itself is designed to address. Specifically, an image that contains a photorealistic AI-generated person and lacks the required metadata tag may be identified by Amazon’s system and flagged for review. The outcome of that review, under the current policy, would be a request to add the metadata or remove the non-compliant image.
The listing suppression consequence
Importantly, listing suppression for an image violation removes the product from search results until a compliant image is uploaded. For a product whose primary or additional images include qualifying AI-generated people without the required tag, suppression would effectively remove the listing from discovery until the images are corrected and re-uploaded with the metadata tag. Notably, the suppression risk is particularly significant for sellers heading into Q4, when listing visibility is at its commercial peak. Addressing non-compliant images before October is the appropriate Q4 preparation step for any seller with Amazon AI generated image assets featuring synthetic performers.
How to Audit Your Existing Catalog
Overall, the audit process has three steps: identifying which images in the catalog contain photorealistic AI-generated people, checking whether those images already have the metadata tag, and adding the tag to any qualifying image that lacks it.
Step 1: Identify AI-generated lifestyle images
First, review the catalog for product images and A+ Content that include lifestyle photography featuring people. Of those, identify which were created using AI generation tools rather than photographed with real models. Any image created using an AI generation platform where the person depicted is synthetic rather than photographed triggers the requirement. Images generated using platforms like Midjourney, DALL-E, Stable Diffusion, or similar tools should be assumed qualifying until reviewed against the exemption criteria.
Step 2: Check existing images for the tag
Second, for images already in Seller Central, download the current file and check whether the contains-synthetic-performer keyword is present in the dc:subject XMP field using ExifTool or another metadata reader. A correctly placed tag means the image is compliant. An absent tag means it requires tagging and re-uploading.
Step 3: Tag and re-upload non-compliant images
Finally, for each non-compliant qualifying image, add the metadata tag using one of the tools described above, verify the tag was written to the correct field, and re-upload the tagged file to Seller Central to replace the current non-compliant version. Amazon moderators recommend completing this retroactive update for all existing qualifying content, not only for new images uploaded after the announcement.
Frequently Asked Questions About the AI Image Metadata Requirement
I used AI to generate a background for a product image, but the product itself is real and there are no people in the image. Does the requirement apply?
No. The requirement applies only to images containing photorealistic AI-generated people. An AI-generated background with no people does not trigger the Amazon AI generated image metadata requirement regardless of how extensively AI was used to create it. The product itself being real and accurately depicted is required by Amazon’s separate accuracy standards, but that is independent of the synthetic performer metadata requirement.
I used a real person as a base reference for an AI-generated model, then modified the appearance significantly. Does the resulting image trigger the requirement?
This is the most ambiguous case in the current guidance. Specifically, Amazon’s exemption covers “real people, even if they have been altered with AI-powered tools,” suggesting that if the base is a real person, AI modification does not convert them into a synthetic performer. However, if the modification is extensive enough that the resulting image is no longer recognizably derived from the real person, the image may cross into synthetic performer territory. When the answer is uncertain, tagging the image is the conservative and safe approach. The metadata tag creates no harm for compliant images; the absence of the tag for non-compliant images creates enforcement risk.
I listed products using AI-generated images before the June 6, 2026 mandatory compliance date. Am I retroactively non-compliant?
Amazon moderators have recommended retroactively updating existing qualifying content. That recommendation implies that existing non-compliant images are not in compliance with the current policy, even though they appeared before the announcement. The practical priority is to identify and tag qualifying images in the catalog now, regardless of when they were uploaded, rather than to wait for an enforcement action to identify which specific images Amazon has flagged.
Does the metadata requirement apply to images in Amazon stores outside the United States?
Amazon states that the requirement applies across its worldwide stores, so sellers should treat it as global rather than limited to the United States. Even though the New York law is what triggered the policy, Amazon’s implementation is global. Qualifying images in any Amazon marketplace worldwide require the metadata tag.
How DAM Law Firm Can Help
DAM Law Firm handles listing compliance issues when AI image policy enforcement produces listing suppression or account consequences, and advises sellers on the compliance steps required to restore listings and avoid recurrence.
Listing reinstatement after image-related suppression
When a listing has been suppressed due to image policy non-compliance and uploading corrected images has not restored the listing, our Amazon listing suspensions team escalates the reinstatement request through Seller Central case escalation with the corrected image documentation. For sellers whose image policy enforcement has escalated to broader account consequences, our Amazon account suspensions team handles the account-level response while listing reinstatement proceeds in parallel.
Compliance review for sellers using AI imagery at scale
For sellers who use AI-generated imagery across large catalogs and want to confirm their compliance posture before enforcement actions surface, our Amazon product compliance team reviews the catalog’s AI-generated image inventory against the requirement and advises on the specific tagging and re-upload process for non-compliant images.
Related DAM Law Firm services:
- Amazon Listing Suspensions: reinstatement after image policy enforcement and listing suppression
- Amazon Account Suspensions: account-level response when image policy non-compliance escalates beyond listing suppression
- Amazon Product Compliance: catalog-level AI image compliance review for sellers using synthetic imagery at scale
This article is for general informational purposes only and does not constitute legal advice. Amazon’s image policies and enforcement practices are subject to change. Contact DAM Law Firm for legal advice tailored to your situation.
Related articles from DAM Law Firm
- Amazon AI Generated Image Metadata 2026: The Complete Technical Guide
- Amazon ASIN Creation Policy Enforcement 2026: What Sellers Must Fix Now
- Amazon Account Health Before Q4: The Pre-Peak Audit Every Seller Needs
- Amazon Product Compliance Lawyer: When Amazon’s Compliance Requirements Need Legal Help
- Amazon Brand Protection 2026: The Legal Framework That Makes Enforcement Actually Work
