Write Amazon’s disclosure keyword into your image metadata

In the dc:subject field of the file's XMP metadata. The official page says to use an IPTC-compatible metadata editor to add the keyword contains-synthetic-performer to the dc:subject (XMP) field of your image file. Many IPTC-compatible editors surface this property under a Keywords label — that is the editors' wording, not Amazon's; the official page names only dc:subject (XMP). This page writes to that same property: the XMP APP1 segment of a JPEG, or the XML:com.adobe.xmp iTXt chunk of a PNG.

contains-synthetic-performer, in lowercase, with two hyphens and no spaces. We write it character for character as it appears in the sentence we captured from the official page; a different spelling would be a different keyword.

The rule covers images that contain photorealistic people made with AI. The official page scopes the keyword to images that contain photorealistic people made with AI — read it for its own wording. Figures that are not photorealistic are not covered; the official page does not list which styles count, so read it before deciding.

Our record paraphrases the official page as naming four cases that do not need the keyword — we hold no verbatim capture of that list, so treat it as our reading of the page rather than its wording: the image shows only real people (even if an AI tool was used to edit them); it shows only characters from films, games or similar works (e.g. a movie poster or game cover); it contains no people at all; or the people in it are not photorealistic. Read the official page for its own wording before deciding.

We cannot say. We hold no captured Amazon sentence about the law in any market, so this page makes no legal claim and gives no legal advice — read the linked official page and check your own market’s rules.

No. The file is not decoded or re-encoded. Pixel dimensions, image quality and file format stay exactly as they were. Only a metadata segment is added or extended, so the file becomes slightly larger.

When the existing XMP packet is well-formed, the keyword is appended to the dc:subject list and the other entries stay. EXIF and ICC segments are kept as they are. If the existing packet cannot be parsed, this page replaces it with a minimal one and the other entries in it are lost.

No. This page writes the keyword into JPEG and PNG files only. The file formats Amazon accepts are listed on Amazon's product image guide and its image technical-requirements page — see the sources below. For any other file type, use an IPTC-compatible metadata editor, which is the method the official page names.

Amazon's own sentence is quoted in full in the spec table above. In short, Amazon says it will add a disclosure where applicable if you include this metadata. We cannot say when or where that disclosure appears. This page only writes the keyword.

We cannot say. The official page carries no publication or revision date that we captured, so we do not give a year. We verified its wording on 2026-09-19; the linked page is authoritative for the current text.

No. The keyword answers one rule only. Background colour, product size in the frame, file naming and file format are separate rules on the same and related official pages, each covered by its own tool below.

Open the downloaded file in any IPTC-compatible metadata editor or XMP viewer and look at the dc:subject property. Many IPTC-compatible editors surface this property under a Keywords label — that is the editors' wording, not Amazon's; the official page names only dc:subject (XMP).