Key Takeaways:
Amazon's new labeling mandate for AI-generated people in product images marks the first major platform-level response to a patchwork of state disclosure laws that threatens to reshape how 2 million third-party sellers create advertising content.
Amazon told third-party sellers on July 23 that any product images or videos containing "AI-generated people" must be tagged with specific metadata keywords before upload, according to a company announcement viewed by CNBC. The policy responds to a New York law that took effect June 9 requiring conspicuous disclosure when advertisements feature "synthetic performers" — digitally created figures meant to read as human — with some exemptions for TV, video game and movie characters.
"The absence of a federal standard means platforms face a compliance puzzle that grows more expensive with each new state law," said Elena Fischer, a regulation analyst at Edgen. "Amazon's move is pragmatic — it builds one system that can scale to multiple jurisdictions rather than retrofitting for each one."
The New York law, which Governor Kathy Hochul described as "first-in-the-nation," applies to "digitally-created media that appear as a real person" distributed to New York audiences. California earlier in 2026 began requiring large AI providers to embed watermarks in AI-generated images, video or other content. The European Union's AI Act brings transparency obligations for AI-generated content into effect in August 2026, with a focus on labeling deepfakes and ensuring generated content is identifiable. No federal law in the United States currently requires companies to disclose when advertising content has been created using AI.
Amazon said it will "add an indicator" to listings informing consumers that images or other content feature AI-generated people "where applicable," though the company did not specify the criteria for when the label will appear. The policy applies to images and "A+ content," which includes videos or other graphics on listing pages. It does not cover content featuring real people even if they have been altered using AI, the company clarified.
The compliance burden falls disproportionately on Amazon's third-party sellers, who account for more than 60 percent of goods sold on its marketplace. Many of these sellers have increasingly used AI tools — including Amazon's own generative AI features — to create text, images and other content for their listings. The new tagging requirement adds a layer of operational cost to a seller ecosystem already navigating listing optimization, advertising fees and inventory management.
The legal stakes extend beyond labeling compliance. A lawsuit filed in New York state court by fashion model Francheska Pujols against retailer Rainbow USA illustrates the liability risk. Pujols alleges that Rainbow used AI to create new depictions of her in poses, body positions and wardrobe that she never modeled, exceeding the scope of a license agreement that permitted use of specified photographs. The complaint asserts claims under New York's Civil Rights Law Sections 50 and 51, Section 43(a) of the Lanham Act for false endorsement, and defamation. The case, which remains at its earliest stages, tests a question no existing talent agreement was drafted to address: when does an authorized modification of a licensed image become an entirely new depiction requiring separate permission?
Meta, TikTok, Pinterest and Google's YouTube have already added AI-generated content labels to videos and images uploaded to their platforms. TikTok and Meta have faced criticism for not adequately labeling ads featuring AI-generated influencers promoting products, in some cases without a brand's knowledge. Amazon's policy goes further by requiring proactive metadata tagging from the seller before upload, shifting the disclosure burden upstream.
The direction of travel is clear. With the EU AI Act's transparency provisions arriving in August 2026 and more US states likely to follow New York and California, platforms face a multi-jurisdictional compliance environment that rewards early standardization. For Amazon's 2 million third-party sellers, the cost of noncompliance — delisted products, legal liability, or consumer trust erosion — now exceeds the cost of building disclosure into the creative workflow. The question is no longer whether AI-generated advertising content gets labeled, but who bears the cost of doing it wrong.
This article is for informational purposes only and does not constitute investment advice.
