If your brand has started using AI to create lifestyle images, models, or product scenes for Amazon, there's a new compliance step you need to know about.
Amazon is now requiring sellers to identify photorealistic AI-generated people in product images and A+ Content before uploading those assets. The requirement uses a specific metadata tag called “contains-synthetic-performer”. Amazon says the change is tied to recent legislation requiring disclosure when advertisements contain photorealistic AI-generated people.
For brands that have been using generative AI to produce listing creative faster, this is more than a technical detail. It changes how your content team needs to prepare and review images before they go live.
What Exactly Is Amazon Asking Sellers to Do?
The basic requirement is pretty straightforward.
If an image or video contains a photorealistic person who was generated by AI, sellers need to add the keyword “contains-synthetic-performer”to the image or video's dc field in its XMP metadata.
Amazon says sellers should use an IPTC-compatible metadata editor to add the tag before uploading qualifying assets to listings or A+ Content. Where applicable, Amazon will then use that information to display an indicator telling customers that the content includes AI-generated people.
So, this isn't a visible watermark that you need to place across the image. It is metadata embedded in the digital file.
That distinction matters because your creative team may not notice anything different when looking at the finished image. The compliance information sits inside the asset itself.
Which Images Actually Need the Tag?
This is where sellers need to pay attention.
Amazon's rule applies to content featuring photorealistic AI-generated people. That can include product images, lifestyle imagery, videos, and A+ Content.
But Amazon also lists situations where the metadata isn't required.
You don't need to add the tag when:
- The image only contains real people, even if AI tools were used to alter their appearance.
- The image contains fictional characters from movies, television, streaming content, documentaries, or video games.
- There are no people in the image.
- The people shown aren't photorealistic.
That makes the first question for your creative team pretty simple: Was the person actually generated by AI, and does that person look photorealistic?
If the answer is yes, treat the asset as a tagging candidate.
Why This Matters for Amazon Seller Management
For teams handling hundreds or thousands of ASINs, this can quickly become a workflow issue.
A brand might have an original product photograph, then use AI to place that product into a lifestyle scene with a completely synthetic model. From a creative perspective, it may look like a normal marketing image. From Amazon's perspective, it may now require additional metadata.
That means Amazon Seller Management can no longer treat creative production and listing compliance as completely separate processes.
Your content workflow should include an AI-content check before assets are uploaded. Every new lifestyle image, product video, and A+ module should be reviewed for synthetic people, especially when multiple AI tools or external creative agencies are involved.
This becomes even more important when you're operating across international marketplaces. Amazon's announcement says the requirement applies to its worldwide stores.
The “Synthetic Performer” Label Is Really a Content Management Issue
The interesting part is that the biggest change may happen behind the scenes.
Brands have become comfortable using AI for faster Content Management Optimization, from generating product scenes to creating alternative creative concepts. Now, metadata has to become part of that workflow.
A practical process could look like this:
Create → Review → Identify AI-generated people → Add metadata → Upload → Verify
That extra review step is small, but it can prevent a team from accidentally uploading an untagged asset.
It's also worth creating a central record of which listing images were generated or substantially created using AI. If your team has hundreds of creative files sitting across folders, agencies, and design platforms, figuring out which assets need tagging later can become much harder.
What Sellers Should Do Right Now
If your brand already uses AI-generated people in Amazon creative, don't wait until the next listing update to address this.
Start with an audit of your existing product images, lifestyle photos, videos, and A+ Content. Identify assets containing photorealistic synthetic people and make sure qualifying files have the required metadata before future uploads.
For new creative, make the AI check part of the production checklist rather than treating it as an afterthought.
And if an external agency creates your listing imagery, add the requirement to the creative brief. Your Amazon Seller Management team should know whether the agency is generating people with AI and who is responsible for embedding the metadata.
For brands working with a Global Ecommerce Accelerator, this is also a good reminder that marketplace growth increasingly involves more than pricing, advertising, and inventory. Creative compliance is becoming part of the operating system.
The same applies to Content Management Optimization. As AI becomes a bigger part of ecommerce content production, teams need systems that track not only what an asset says or looks like, but also how it was created.
For a Global Ecommerce Accelerator, that means AI governance should sit alongside listing management, content workflows, and marketplace compliance rather than living in a separate silo.
And for brands expanding across marketplaces, the lesson is pretty simple: AI can make creative production faster, but faster content still needs the right compliance checks before it reaches customers.
One more thing to watch: Amazon's current guidance is specifically focused on photorealistic AI-generated people. So don't assume every image created or edited with AI needs this tag. The trigger is the presence of a qualifying synthetic person, not simply the fact that an AI tool touched the file.