- Amazon told sellers on July 22 to identify product media containing photorealistic people generated entirely by AI.
[1] - Affected images must carry the keyword
contains-synthetic-performerin the file’s XMP metadata; Amazon will display a disclosure where applicable. [2] - New York’s law carries civil penalties of $1,000 for a first violation and $5,000 for subsequent violations, but exempts platforms that merely publish or distribute the ads.
[3] Amazon is requiring third-party sellers to identify product images containing photorealistic, AI-generated people, introducing a machine-readable labeling process after a New York advertising disclosure law took effect. The company notified sellers on Wednesday, July 22.[1]
Under Amazon’s product image guide, sellers must use an IPTC-compatible editor to add the keyword contains-synthetic-performer to the dc:subject field in an image’s XMP metadata before up it. Amazon says it will use that tag to show shoppers a disclosure where applicable.[2]
A narrow rule for synthetic people #
Amazon’s requirement covers photorealistic people created entirely by AI and not based on a real person. Sellers should not apply the tag to images that contain only real people, including people altered with AI tools; images without people; non-photorealistic characters; or characters from movies, games and similar expressive works.[2]
The guide does not disclose a separate grace period or a specific penalty for missing AI metadata. Amazon’s general image rules allow it to remove images that violate selling policies and suppress a listing from search when it lacks a compliant main image.[2]
New York targets the advertiser, not the platform #
New York General Business Law Section 396-b took effect June 9. It requires a business that produces or creates a commercial advertisement—and has actual knowledge that it contains a synthetic performer—to make a conspicuous disclosure. Violations carry a $1,000 civil penalty for the first offense and $5,000 for later offenses.[3] The statute exempts media that merely publish or disseminate an offending ad, including streaming services and other advertising platforms. Amazon’s policy nevertheless shifts the initial identification step to sellers while letting the marketplace control when and where shoppers see the label.[2][3]
Early seller discussion has focused on implementation. In Amazon’s seller forum, merchants asked whether the requirement applies to previously uploaded images as well as new submissions, a point the public guide does not expressly resolve.[4]
Companies mentioned #
Further sources #
[[1] CNBC, “Amazon makes sellers label AI-generated people in images after NY law,” … ↗](https://www.cnbc.com/2026/07/23/amazon-makes-sellers-label-ai-generated-people-in-images-after-ny-law.html)
[[2] Amazon Seller Central, “Product image guide,” AI-generated people disclosure an… ↗](https://sellercentral.amazon.fr/help/hub/reference/external/G1881?locale=fr-FR)
[[3] New York State Senate, S8420-A, amendment to General Business Law Section 396-b. ↗](https://www.nysenate.gov/legislation/bills/2025/S8420/amendment/A)
[[4] Amazon Seller Forums, “What new rules about AI-generated product images should … ↗](https://sellercentral.amazon.com/seller-forums/discussions/t/33503429-9a8a-4925-bef6-944b4c8ea1ac)
The stories that matter, in one email. Free — unsubscribe anytime.