Amazon and AI-Generated Visuals: How to Use the contains-synthetic-performer Tag

Amazon now requires sellers to identify images and videos featuring fully AI-generated, photorealistic human performers. Before uploading any media assets, you must insert the keyword contains-synthetic-performer into the dc:subject XMP metadata field.

This requirement does not apply to all AI-generated or edited content. It targets a highly specific use case: media depicting a synthetic individual designed to look real, but not modeled on a real, identifiable person.

This metadata enables Amazon to flag relevant assets and comply with transparency regulations in various jurisdictions. It does not replace Amazon’s existing compliance rules for product images, video content, or A+ Content.

Key Takeaways

• The tag applies only to fully AI-generated, photorealistic human models—not to all AI-generated brand assets.

• The exact keyword required is contains-synthetic-performer.

• The keyword must be saved in the file's dc:subject XMP field.

• Real human models edited using AI tools do not require this specific tag.

• This policy is explicitly outlined in Amazon’s product image, A+ Content, and video guidelines.

• Amazon has not documented any impact of this metadata on product rankings, conversion rates, or search visibility.

What Exactly Does Amazon Require?

Amazon specifies that affected media assets must include the following value:

contains-synthetic-performer

This value must be added to the dc:subject field within the file's XMP metadata. Specifically, Amazon requires the keyword to be saved as a value within the rdf:Bag XMP container.

Using an IPTC/XMP-compatible tool allows you to inject this metadata without altering the visual presentation of the asset.

This tag is not a visible watermark on the image. It is metadata embedded directly within the file structure, which Amazon reads during the upload process.

Amazon outlines the exact workflow in its official help center under guidelines for tagging AI-generated human performers in media.

In Vendor Central, this documentation can be located via:

Help → Images & Video → Media Tagging for AI-Generated Human Performers

Which Creative Assets Require the contains-synthetic-performer Tag?

The tag is mandatory if an image or video contains at least one fully AI-generated human model presented in a photorealistic style.

These specific scenarios and exclusions are detailed in the official Amazon Product Image Guidelines.

Note: Being exempt from this metadata tag does not release brands from other requirements regarding copyright, model release rights, trademark compliance, or accurate product representation.

What About Digital Twins of Real People?

Digital twins or replicas representing identifiable real people must be managed differently.

They do not fall neatly into the definition of a synthetic performer (which refers to fully AI-generated individuals not modeled on real people). However, this does not mean their use is automatically authorized.

A technical metadata tag cannot resolve legal questions about consent, likeness rights, identity theft, or advertising claims. In case of doubt, legal validation is required, not just technical compliance.

Where Does This Policy Apply on Amazon?

Verified official documentation lists several media categories:

• Product detail page gallery images;

• Images embedded in A+ Content;

• Videos embedded in A+ Content;

• Other product videos governed by applicable Amazon guidelines.

Both the Amazon A+ Content Guidelines and the Product Video Guidelines mandate using the contains-synthetic-performer keyword in the dc:subject XMP field.

This documentation is currently live across several Seller Central domains in Europe, North America, and other regions.

However, Amazon has not published a single global enforcement date or market-by-market rollout timeline. Brands should verify the guidelines in local Seller/Vendor Central accounts before any major catalog update.

How to Add the Metadata on macOS

Amazon recommends using the native Preview app to write the keyword metadata:

1. Open the final asset in Preview.

2. Select Tools, then Show Inspector, or use Cmd + I.

3. Go to the Keywords tab.

4. Click the + button.

5. Type exactly contains-synthetic-performer.

6. Press Enter.

7. Save the file.

Apple confirms in its support documentation that Preview supports assigning keywords to image files.

For enterprise workflows, we recommend validating the XMP field using a secondary tool prior to Amazon ingestion.

How to Add the Metadata on Windows

Amazon’s suggested manual workflow uses File Explorer:

1. Right-click on the final image or video file.

2. Select Properties.

3. Go to the Details tab.

4. Locate the Tags field.

5. Add contains-synthetic-performer.

6. Apply and save changes.

This manual method depends heavily on the file format and how Windows handles its specific metadata properties.

After saving, verify that the value is indeed mapped to the dc:subject XMP field and not written to a proprietary Windows-only field that Amazon's systems cannot read.

How to Automate Metadata Tagging with ExifTool

For digital asset management teams handling high volumes of creative assets, automation can be set up via ExifTool, a command-line application used to read and write metadata across many file formats.

To append the value without overwriting existing keywords:

exiftool -XMP-dc:Subject+=contains-synthetic-performer image.png

To verify the field content after processing:

exiftool -XMP-dc:Subject image.png

The console output should display:

contains-synthetic-performer

The += operator is critical: it appends the new value to the existing array rather than wiping out other essential XMP keywords.

For video assets, metadata retention depends heavily on container formats and export workflows. Always audit final video outputs post-export.

Common Pitfalls to Avoid

Over-Tagging All AI-Generated Content

AI-generated backgrounds, packshots, or abstract illustrations do not automatically trigger this requirement.

Amazon is specifically looking for photorealistic synthetic human models.

Adding Metadata Prior to Final Export

Image compression, format conversions, resizing, or subsequent exports often strip out metadata.

The metadata tag must be applied to or verified on the final, production-ready file uploaded to Amazon.

Confusing XMP, EXIF, and Comments Fields

Amazon specifically scans the dc:subject XMP field.

Even though ExifTool is the standard utility for editing these tags, simply putting the keyword into any generic EXIF field will fail Amazon's compliance checks.

Overwriting Existing Asset Keywords

In automated workflows, ensure you append the tag to the existing list rather than overwriting other valuable product metadata.

Keep a master copy of the source asset and maintain an audit log of modified metadata.

Assuming Tagging Guarantees Asset Compliance

Transparency tags do not validate misleading content.

Product representations must remain structurally accurate. AI assets must not fabricate product features, dimensions, performance, or use cases that the physical product does not offer.

Expecting Direct Impact on SEO or Best Seller Rank

Amazon's official documentation frames this metadata strictly as a transparency and regulatory compliance measure.

There is no indication of ranking penalties or boosts associated with the tag. Any claims regarding impact on organic ranking, ad performance, or algorithmic recommendation are speculative.

Why is Amazon Enforcing This Tag?

The integration of AI-generated models in brand marketing is scaling fast. Synthetic models are widely used in lifestyle imagery, product demos, videos, and AI-driven UGC-style content.

Amazon states that this tagging helps satisfy regional legal frameworks that mandate clear disclosure when synthetic humans are used in commercial advertising.

This aligns with broader regulatory shifts. For example, New York State has enforced transparency laws regarding synthetic performers in commercial contexts since June 9, 2026. The official legislative text outlines that platforms and brands must be able to trace the provenance of human likenesses in creative assets.

While the rule applies globally across Amazon marketplaces, this legal backdrop explains why Amazon is structuring metadata ingestion from brand partners now.

What This Means for Brands & Digital Asset Management

The quality and compliance of PDP content are no longer determined solely by what is visible on the screen.

E-commerce content and production teams must now track:

• Asset provenance and origin;

• Production workflows and tools used;

• Whether depicted models are real or synthetic;

• Associated usage rights and model releases;

• Embedded metadata structure within files;

• Post-processing transformations prior to upload;

• Target marketplaces where the assets will be active.

Brands must be able to answer several key questions for every creative asset:

• Who produced the asset?

• What software or AI generator was used?

• Is the person represented real or synthetic?

• Is the style photorealistic?

• Are real and synthetic models mixed in the creative?

• Are intellectual property rights and compliance approvals documented?

• Is the metadata intact in the final production file?

• Did an external agency or asset compression tool strip the tags?

• Is this identical asset distributed across global marketplaces?

Action Plan for E-commerce Teams

1. Audit Active Assets. Document all live and pending product images, A+ Content, and product videos.

2. Classify Creative Content. Categorize assets by human models (real vs. real with AI edits vs. fully AI-generated), non-photorealistic illustrations, and assets with no human presence.

3. Isolate Affected Files. Identify the specific images and videos containing at least one photorealistic, fully AI-generated performer.

4. Review Licensing Rights. Confirm commercial usage rights, model releases, and brand compliance for all generated assets.

5. Inject Tag at Final Export. Ensure the contains-synthetic-performer tag is written to the file after all final compression, resizing, and exporting.

6. Validate XMP Fields. Perform metadata audits to verify that the tag resides in the dc:subject property.

7. Maintain an Asset Log. Archive the original source file, final uploaded file, timestamp, target marketplace, and compliance validation.

8. Audit the Live PDP. Verify that published assets render correctly without any unexpected changes.

9. Monitor Amazon Policy Updates. Keep abreast of evolving regional guidelines, marketplace-specific programs, and emerging media requirements.

Conclusion

Amazon does not require disclosure for every application of AI in asset generation. The contains-synthetic-performer rule targets photorealistic human models generated entirely by AI.

While the technical file modification is straightforward, the operational challenge lies in asset governance: accurately identifying affected assets, preserving metadata throughout the content supply chain, validating rights, and preventing third-party agencies or optimization tools from stripping out compliant tags.

Detail page compliance now depends as much on the invisible metadata embedded within your brand assets as it does on the visual content itself.

Frequently Asked Questions

Does Amazon require this tag for every AI-generated image?

No. The contains-synthetic-performer tag applies only to assets featuring fully AI-generated, photorealistic human models. Creative assets without human figures or those featuring non-photorealistic characters are out of scope.

Do I need to tag real photos edited with AI?

No, not under this specific policy, provided the models are real people—even if AI was used for background replacement or retouching. General compliance and advertising accuracy standards still apply.

Does this policy apply to A+ Content?

Yes. Amazon's official A+ Content Guidelines state that relevant assets must carry the contains-synthetic-performer tag within the dc:subject XMP field.

Does the tag need to be visually displayed on the image?

No. Amazon requires this to be file-level metadata. Adding a visible text disclaimer on the image itself does not satisfy the technical compliance requirement.

Will using this tag hurt my product's search visibility?

There is no documentation showing that this tag affects search rankings or advertising performance positively or negatively. Its sole purpose is legal compliance and platform transparency.

What is the difference between XMP and EXIF for this tag?

XMP and EXIF are distinct metadata schemas. Amazon explicitly scans the XMP dc:subject property. While utilities like ExifTool can write to this field, writing the tag to generic EXIF parameters will not pass Amazon's verification.