Amazon AI Product Image Compliance 2026: Synthetic Performer Metadata Tagging Guide

TLDR. On July 27, 2026, Amazon began requiring file-level metadata tags for any product image containing photorealistic AI-generated people. You must add the keyword contains-synthetic-performer to the IPTC/XMP dc:subject field before upload using tools like Adobe Photoshop, Bridge, ExifTool, or PixPix. Images with real people (even if AI-retouched) and product-only shots do not require the tag. Failure to comply triggers listing suppression and repeat violations can escalate to account suspension. This rule creates unmet demand in product categories where authentic photography becomes a competitive moat because competitors rely on AI model imagery.

What is Amazon’s synthetic performer metadata rule?

Amazon now requires that any product image, A+ content module, or product video showing a photorealistic AI-generated person carry the IPTC/XMP keyword contains-synthetic-performer in the dc:subject metadata field at the file level. The rule went live on July 27, 2026, and enforcement began on August 15, 2026. This is not a visible watermark. The tag lives in the file metadata and Amazon’s systems read it during upload to determine whether to show a disclosure indicator to shoppers.

The policy applies globally. It covers main images, secondary images, lifestyle shots, A+ content, brand stores, and product videos. It does not apply to images of real people (even if you used AI to remove a background or adjust lighting), non-photorealistic illustrations, or images without people. It only applies when you generate a realistic human who does not exist.

Amazon ties this rule to three recent laws: New York’s Synthetic Performer Disclosure Law (operative June 2026), the EU Artificial Intelligence Act (rollout August 2, 2026), and California’s AI Transparency Act (operative August 2, 2026). All three require disclosure when ads or product materials depict synthetic humans. Amazon operationalizes these frameworks with a single metadata schema sellers can apply at scale.

Why this matters now: multiple seller blogs and forums in late August 2026 report that Amazon’s AI detection systems are suppressing listings that violate main image rules and tagging patterns. Sellers describe “mass bans,” “zero-warning suppression,” and opaque enforcement. The metadata tag is the signal that tells Amazon’s system you disclosed properly. Missing it can trigger automated removal and repeated violations escalate to account suspension.

How to add the contains-synthetic-performer tag to your images

You write the tag into IPTC/XMP metadata using an editor that supports the Dublin Core subject field. The tag must be present in the file you upload to Amazon, not just in your working file. Here is the step-by-step workflow.

Step 1: Export all active ASIN images and classify them

Pull every image and video for your active listings. Sort them into three buckets:

1. Images with human figures.
2. Images where those humans are AI-generated (fully synthetic, not real people edited with AI tools).
3. Images where you still have the source file with editable metadata.

If you used a tool like Midjourney, DALL-E, or Stable Diffusion to create a model or customer in a lifestyle scene, that image needs the tag. If you photographed a real person and used Photoshop’s generative fill to clean the background, that image does not need the tag. If you used an AI tool to create a product render with no people, no tag is required.

Step 2: Choose your IPTC/XMP editor

You need software that can batch-edit IPTC metadata. The most common tools are:

• Adobe Photoshop: Open the image, go to File → File Info, select the IPTC tab, find the Keywords field (which maps to dc:subject), add contains-synthetic-performer, and save.
• Adobe Bridge: Select multiple images, open File Info, add the keyword in the IPTC tab, and apply to all selected files.
• Adobe Lightroom: Use the Keyword panel to add contains-synthetic-performer to the selected images, then export.
• ExifTool (command line): Run exiftool -IPTC:Keywords+=contains-synthetic-performer image.jpg to append the keyword. Use a shell script to batch-process folders.
• PixPix (Windows/Mac GUI): Open the image, navigate to the IPTC section, add the keyword, and save.

For large catalogs (hundreds or thousands of images), ExifTool or Adobe Bridge are the fastest. ExifTool can process entire directories in seconds and log each change for audit trails.

Step 3: Write the keyword to dc:subject and verify

The IPTC Keywords field in most editors corresponds to the XMP dc:subject array. Add contains-synthetic-performer exactly as written (lowercase, hyphenated, no spaces). Do not add extra keywords like “AI-generated” or “synthetic” unless you want them for internal DAM purposes. Amazon looks for the exact string.

After tagging, verify the metadata using ExifTool or Bridge. Open the file and confirm the keyword appears in the IPTC section. If you export from Lightroom or Photoshop, re-check the exported file. Some workflows strip metadata during export if settings are misconfigured.

Step 4: Upload the tagged file to Amazon

Upload the image through Seller Central, Manage Inventory, or bulk upload templates. Amazon’s system scans the file metadata on upload. If the tag is present and correct, the image is accepted and a disclosure indicator is added to the listing for shoppers. If the tag is missing on an image that Amazon’s AI detection flags as containing a synthetic person, the listing can be suppressed.

Step 5: Store compliance records

Keep a log of which ASINs received tagged images, the date of upload, and a copy of the tagged source file. If Amazon challenges a listing, you need proof that you disclosed properly. Export the IPTC data from your DAM or use ExifTool to generate a CSV report of all files with contains-synthetic-performer in their metadata.

When you do NOT need the synthetic performer tag

Amazon’s guidance is narrow. You only tag photorealistic AI-generated humans. You do not tag:

• Images of real people, even if you used AI to edit exposure, remove blemishes, replace backgrounds, or adjust colors.
• Non-photorealistic characters from films, TV shows, video games, or illustrated art.
• Product-only images with no people (even if the product itself was AI-generated).
• Images where the AI tool generated props, backgrounds, or abstract elements but no human figures.
• Hands, arms, or partial body shots of real people, even if AI-retouched.

If you are unsure whether your image depicts a “photorealistic AI-generated person,” ask: could a shopper mistake this for a photo of a real human? If yes and the human is fully synthetic, tag it. If the base image is a photograph of a real person and you only used AI for cleanup or composition, do not tag it.

Multiple seller education posts from August 2026 stress this distinction because early confusion led to over-tagging. Over-tagging is not a violation, but it adds unnecessary disclosure indicators that may reduce conversion. Under-tagging is the compliance risk.

What happens if you do not comply

Amazon’s enforcement is automated and escalates with repeat violations. Here is what sellers report in late August 2026:

1. First violation: Listing suppression. The ASIN is removed from search and buy box. You receive a notice to fix the image. Once you upload a compliant version, the listing is restored, usually within 24 to 48 hours.
2. Repeat violations: If Amazon flags three or more high-risk violations in 90 days (including synthetic performer tags, main image spec issues, and misleading imagery), your account enters a compliance review. You may be required to submit all image source files and metadata logs.
3. Egregious or sustained violations: Account suspension. Amazon can suspend selling privileges for 30 days or permanently for sellers who repeatedly upload undisclosed AI-generated people or intentionally misleading product images.

Seller forums describe these escalations as “mass bans” and “AI purge” because the detection is AI-driven and batch enforcement waves hit multiple sellers simultaneously. The system is opaque. You do not get a warning before the first suppression, and Amazon does not provide a detailed explanation of why an image was flagged. The burden is on you to audit and correct.

This is where Amazon AI image compliance checklists become critical. You need to run pre-upload scans that verify metadata tags, background purity, aspect ratios, and product visibility before the file reaches Amazon’s servers.

Why this rule creates unmet demand opportunities

Amazon’s synthetic performer disclosure rule and the broader main image enforcement wave in August 2026 reveal a structural arbitrage. Product categories that rely heavily on AI-generated model imagery now carry compliance friction and shopper trust risk. Categories where authentic photography is easy and affordable become relatively more attractive.

Here is the logic:

1. Listings with the contains-synthetic-performer tag show a disclosure indicator. Amazon describes this as labeling AI-generated images “less trustworthy.” Early conversion data is anecdotal, but operators on X and LinkedIn report lower click-through and higher bounce on listings with visible AI labels in apparel, beauty, and home decor.
2. Competitors who built their imagery stack on AI models (to save costs on photoshoots) now face a choice: accept the trust penalty or reinvest in real photography.
3. New entrants who budget for authentic photography from day one have a moat. Their listings do not carry the synthetic performer tag, they avoid the compliance audit risk, and they present as more established brands.

This is most pronounced in categories where lifestyle imagery drives purchase decisions: apparel, accessories, beauty, pet products, home textiles, and outdoor gear. It is less relevant in categories where product-only shots dominate: electronics components, industrial supplies, raw materials, and tools.

Kaldon’s Discover phase flags this as an authenticity filter when scanning demand gaps. If you see high search volume, weak visual branding among top sellers, and heavy reliance on stock or AI model imagery, that is a signal that authentic photography can become a differentiation lever. You are not just selling a better product. You are selling a more trustworthy presentation in a category where competitors cut corners on creative.

The inverse is also true. If you are evaluating a product category where competitors have invested in high-quality real photography and you plan to use AI models to save costs, factor in the compliance overhead (metadata tagging, audit risk, disclosure labels) and the potential conversion drag. The cost savings from AI imagery may not materialize if your click-through drops 10 to 15 percent because of the synthetic performer tag.

How to integrate metadata tagging into your creative workflow

If you use AI-generated people in your product imagery, metadata tagging must become a standard step in your creative pipeline. Here is how operators are implementing this in August 2026:

Centralized tagging at export

If you use a DAM or content management system, configure it to auto-append contains-synthetic-performer to any file flagged as containing AI-generated humans. This requires a classification step (manual or automated) before export, but it ensures consistency. When the asset is pulled for Amazon upload, the tag is already present.

Pre-upload compliance scans

Run automated scans before files reach Amazon. Tools like BanProof AI and compliance modules in eCommerce intelligence platforms check for the presence of the synthetic performer keyword in IPTC metadata and flag missing tags. This catches errors before suppression.

Quarterly audits of live listings

Amazon’s rules and detection systems evolve. Images that were compliant in Q2 may be flagged in Q3 if Amazon refines its AI detection model. Schedule quarterly audits where you re-export all active listing images, verify metadata tags, and check for new enforcement patterns reported in seller communities.

Source file version control

Keep the original AI-generated file and the tagged, exported version in separate folders with clear naming conventions. If Amazon challenges a listing, you need to prove that the uploaded file contained the tag. Version control and file hashes provide that audit trail.

Cross-platform consistency

If you sell on Amazon, Walmart, Shopify, and run paid ads on Meta or TikTok, you may face different AI disclosure rules. The contains-synthetic-performer tag satisfies Amazon and aligns with IPTC standards used by C2PA (the Content Authenticity Initiative). For visible labels required by the EU AI Act or TikTok Shop, you may need to add overlay text or use platform-specific toggles. Centralizing metadata tagging at the DAM level ensures you only edit once and syndicate correctly.

How Kaldon handles synthetic performer compliance

Kaldon’s Create phase integrates metadata tagging into the asset generation workflow. When you generate lifestyle or model imagery using AI, the system detects the presence of photorealistic humans and automatically writes contains-synthetic-performer to the IPTC dc:subject field before export. You do not need to remember to tag manually. The compliance step is built into the tool.

Kaldon also flags unmet demand in categories where authentic photography is a competitive advantage. During the Discover phase, the platform scans top listings for reliance on stock imagery, AI-generated models, or low-quality visuals. If a category shows high search volume, weak visual branding, and heavy use of synthetic imagery, Kaldon surfaces it as an opportunity where investing in real photography (or using compliant, high-quality AI renders) can drive differentiation.

This is the arbitrage Amazon’s new rule creates. Compliance friction becomes a moat. Sellers who operationalize metadata tagging and invest in authentic imagery gain trust, avoid suppression risk, and capture demand that competitors lose to enforcement waves.

For sellers managing large catalogs, Kaldon’s batch audit tools scan all active ASINs, detect images with synthetic performers, verify the presence of the metadata tag, and generate compliance reports for Amazon appeals or internal audits. This is critical in August 2026 as enforcement escalates and sellers report “zero-warning suppression” for missing tags.

Start your free trial to see how Kaldon automates synthetic performer metadata tagging and flags AI-resistant product opportunities.

Broader context: main image compliance beyond synthetic performers

The synthetic performer tag is one piece of Amazon’s August 2026 enforcement wave. Sellers are also being hit with suppression for main image violations that have nothing to do with AI. The most common patterns are:

1. Non-white backgrounds: Main images must have a pure white background (RGB 255, 255, 255) in most categories. AI-generated images with off-white, gradient, or textured backgrounds are being flagged.
2. Product visibility under 85 percent: Main images must show the product filling at least 85 percent of the frame. Lifestyle shots with excessive white space or context are being suppressed.
3. Composite images: Combining multiple product angles or views in a single main image is prohibited in most categories. This rule applies whether the composite is AI-generated or manually assembled in Photoshop.
4. Text, logos, or graphics: Main images cannot include promotional text, brand logos (except when embossed on the product itself), or graphic overlays. AI tools that add “Sale” or “New” badges are generating violations.

These rules apply to all listings, not just those with AI-generated people. Sellers in August 2026 are running dual audits: one for synthetic performer metadata and one for general main image spec compliance. Amazon’s full AI image labeling policy covers both angles.

The strategic takeaway: if you are launching a product in a category where competitors are getting hit with suppression for low-quality or non-compliant imagery, that is a signal of unmet demand. Shoppers are searching, but listings are unstable. New entrants with compliant, high-quality images can capture share during the enforcement window.

Practical checklist for immediate compliance

If you need to comply with Amazon’s synthetic performer rule today, follow this sequence:

1. Audit your catalog: Export all images for active ASINs and filter for those containing human figures. Flag any that are AI-generated.
2. Install an IPTC editor: If you do not have Adobe Creative Cloud, download ExifTool (free, command line) or PixPix (GUI, $20 one-time). For large catalogs, ExifTool is faster.
3. Batch-tag synthetic performer images: Add contains-synthetic-performer to the IPTC Keywords field. Verify the tag is present using ExifTool or Bridge.
4. Re-upload tagged images: Replace the existing images in Seller Central with the tagged versions. Do this before Amazon flags the listing.
5. Document compliance: Export a list of all tagged files with timestamps and store it in your DAM or a compliance folder. If Amazon challenges a listing, this is your proof.
6. Set up pre-upload scanning: Add a step to your creative workflow that checks for the presence of the tag before any image is uploaded to Amazon. This prevents future violations.
7. Monitor enforcement: Join seller communities where operators share suppression notices and enforcement patterns. Amazon’s detection model evolves, and new patterns emerge every quarter.

For a more detailed compliance framework, see the Amazon AI image compliance checklist, which covers metadata tagging, main image spec, and cross-platform disclosure rules.

Why this is not just an Amazon problem

Amazon’s synthetic performer rule is the most publicized, but it is part of a broader regulatory shift. New York, the EU, and California now require disclosure of AI-generated people in advertising and eCommerce. TikTok Shop has similar toggles for AI content. Meta and Google are rolling out ad transparency features that flag synthetic humans.

If you sell on multiple platforms or run paid media, you need a unified approach to AI disclosure. The contains-synthetic-performer IPTC tag is a good foundation because it is platform-agnostic and aligns with C2PA standards. You write it once in your DAM and it travels with the file to Amazon, Walmart, Shopify, and any other platform that reads IPTC metadata.

The bigger opportunity is using this shift to filter product ideas. Categories where competitors rely on AI models are now higher-friction. Categories where authentic photography is affordable are lower-friction. Kaldon’s Discover phase uses this as a scoring input when ranking demand gaps. Compliance friction is a form of competitive moat, and right now, authentic imagery is the moat.

Start your free trial to see how Kaldon automates synthetic performer compliance and flags unmet demand in AI-resistant product categories.

Frequently asked questions

What is the exact keyword I need to add to my image metadata for Amazon compliance?

Add contains-synthetic-performer (lowercase, hyphenated, no spaces) to the IPTC Keywords field, which maps to the XMP dc:subject array. Use tools like Adobe Photoshop, Bridge, Lightroom, or ExifTool to write the tag before uploading to Amazon.

Do I need to tag images of real people that I edited with AI tools?

No. The tag is only required for photorealistic AI-generated humans (fully synthetic people who do not exist). If you photographed a real person and used AI for background removal, color correction, or retouching, no tag is required.

What happens if I forget to tag an image with a synthetic performer?

Amazon’s automated detection may suppress the listing. You will receive a notice to correct the image. Repeated violations (three or more in 90 days) can escalate to account suspension. Always verify metadata tags before upload.

Can I use ExifTool to batch-tag hundreds of images at once?

Yes. Run exiftool -IPTC:Keywords+=contains-synthetic-performer *.jpg in a terminal to tag all JPEG files in a folder. Use a shell script to process multiple directories and log each change for audit trails.

How does this rule create product opportunity signals?

Categories where competitors rely on AI models now carry compliance friction and trust penalties (Amazon labels these images as less trustworthy). If you invest in authentic photography, you avoid the disclosure tag and gain a competitive moat in conversion and brand perception.

Sources & citations

• https://ads.tiktok.com/resources/help/article/about-ads-metrics-in-seller-center?lang
• https://ads.tiktok.com/resources/help/article/about-gross-revenue-for-tiktok-shop-ads?lang=en
• https://www.reddit.com/r/AmazonFBA/comments/1vxozcy/product_research_help/
• https://www.reddit.com/r/indianstartups/comments/1vqqnev/i_am_launching_category_intelligence_for_indian/
• https://ads.tiktok.com/resources/help/article/best-practices-for-setting-a-live-shopping-ads-budget?lang=es-419
• https://reddifier.com/free-subreddit-analysis-tool/r/FulfillmentByAmazon
• https://www.directtoconsumer.co/newsletter/reddit-citations-google-ai-video-meta-launch-strategy
• https://seller-ie.tiktok.com/university/essay?knowledge_id=1898073370461966
• https://peerpush.com/p/ideafast
• https://ads.tiktok.com/resources/help/article/about-tiktok-shop-ads-attribution?lang=en
• https://longbridge.com/news/295362908
• https://marketplacemax.com/insights/amazon-ai-generated-people-image-disclosure-2026/
• https://mrjeffamz.com/blog/amazon-seller-news-2026
• https://channelx.world/2026/08/amazon-ai-generated-people-image-and-video-rules/
• https://www.rewarx.com/blogs/amazon-ai-rules-august-2026-trigger-seller-bans-3-high-risk-main-image-patterns