Amazon Mandates AI Disclosure Labels for Synthetic People in Product Images

Amazon has begun requiring third-party sellers to identify and label product images containing photorealistic AI-generated people, a policy shift that transforms AI disclosure from optional to mandatory across the marketplace.

The company notified sellers this week that qualifying images and enhanced A+ content modules must carry specific metadata keywords before upload, according to CNBC, which first reported the details. Amazon plans to add a shopper-facing indicator to listings where applicable, though the exact criteria for when that label appears have not been made public.

Sellers must embed an IPTC-compatible metadata keyword flagging synthetic performers in qualifying assets, eWeek reported. The requirement excludes two categories: fictional characters from film, television, or video games, and real people whose appearance has been altered with AI tools.

Why it matters

This policy puts a compliance decision inside every creative workflow that uses AI imagery. For the 60% of Amazon units sold by third-party sellers, many of whom have adopted AI lifestyle imagery for its speed and cost advantages, the disclosure requirement now sits alongside the efficiency case. Research shows AI-labeled content typically receives lower consumer trust and purchase intent, making this a brand positioning question as much as a legal one.

The law driving the change

New York's synthetic performer disclosure law took effect June 9, 2026, after Governor Kathy Hochul signed the legislation in December 2025. The statute amends General Business Law § 396-b and applies to advertisements across all media.

The law defines a synthetic performer as a digital asset created or modified by generative AI that gives the impression of a human performance by someone not recognizable as an identifiable person. Anyone producing a covered advertisement who knows it contains a synthetic performer must conspicuously disclose that fact. Civil penalties start at $1,000 for a first violation and climb to $5,000 for subsequent ones.

The definition explains why Amazon's carve-outs exist. Digital replicas of identifiable real people fall under separate right-of-publicity laws, and advertisements for expressive works are excepted when the synthetic performer matches the underlying content.

Publishers and distributors put on notice of an undisclosed synthetic performer have five days to remove the ad or bring it into compliance, according to attorneys at Reed Smith. Background figures count—a synthetic bystander in a lifestyle scene triggers the requirement just as a hero model does.

The performance paradox

Amazon actively markets generative AI creative tools to the same sellers now responsible for tagging. The company reports that Sponsored Brands campaigns using AI-generated images delivered 10.3% higher return on ad spend than campaigns without them, based on internal U.S. measurements across 2024 and 2025. Brands using its AI creative tools advertised five times as many products, and advertisers who adopted its image generator saw 10% more sales per advertiser monthly.

Yet available research points to a label penalty. In a study published in Equilibrium in December 2025, researcher Khalil Israfilzade found that advertisements labeled as AI-generated drew lower consumer trust and purchase intent than those labeled human-made, with the gap widest for high-involvement products like laptops.

Separate research published in PNAS Nexus found that labeling headlines as AI-generated lowered perceived accuracy and willingness to share regardless of whether the content was true or human-written, though the effect was roughly a third the size of labeling something false.

What sellers must do

The practical workflow includes five steps: audit catalogs to identify which listings use synthetic human figures, move metadata tagging upstream into asset production, maintain creative records documenting intent, choose creative approaches strategically by category, and test different creative routes through Amazon's experiment tools.

Hybrid production offers a path forward. AI-generated backgrounds behind a real model with proper releases sit outside the synthetic performer trigger, making combined approaches viable rather than compromise solutions.

The regulatory landscape

New York represents the beginning, not the endpoint. TikTok Shop directs sellers to label content substantially created or altered by AI. YouTube began automatically applying AI labels in May when its systems detect significant photorealistic AI use. California's AI Transparency Act becomes operative August 2, with large platform duties beginning January 1, 2027, including a prohibition on knowingly stripping compliant provenance data from uploaded content.

Tech Policy Press counts 109 state AI laws enacted through July 1, 2026. No federal statute currently preempts state disclosure requirements, and advisors generally recommend continued compliance with state law while federal questions are litigated.

The details were first reported by CNBC and eWeek.