Skip to content
Adrythm
AI search

AI content disclosure

labeling AI images / AI generated content disclosure / DigitalSourceType / do we have to say it was AI

In short

AI content disclosure is labeling material as machine-made where a platform requires it. Google's product listings require AI-generated images to carry a specific IPTC metadata code, and IPTC defines three different codes: created with generative AI, edited with it, and edited by a person.

Most guidance about AI content is advice. This one is a requirement with a named standard behind it, which makes it checkable.

Google says generative AI is useful for research and for adding structure to original content. It also says that using such tools to generate many pages without adding value for users may violate its spam policy on scaled content abuse. Then it gives sellers something exact. For Merchant Center, AI-generated images must carry the IPTC DigitalSourceType TrainedAlgorithmicMedia metadata. AI-generated product data such as titles and descriptions must be listed separately and labeled as AI-generated.

IPTC defines that code, and its wording is where the useful part sits. It names the code Created using Generative AI. The definition is digital media created algorithmically using an AI model trained on captured content. In plain terms, a wholly machine-made image.

A real photo with a generative background swap is not that. IPTC gives it a separate code, named Edited using Generative AI. Its definition covers augmentation, correction or enhancement using a generative AI model, such as inpainting or outpainting. A third code, Human-edited media, covers the same kinds of change made by people using non-generative tools.

So the question is not whether AI touched the picture. It is which of three cases applies, and each has its own code that travels inside the file.

In practice

The case most often labeled wrongly is the common one. A genuine product photo has its background replaced, or an object removed, using generative fill. That is not a created image, and it is not a human edit either. Whoever shoots your product photography needs to know the difference, because the code is written then and nobody downstream can work it out later.

Not the same as

Scaled content abuse
That is a spam policy about mass production without value. This is a labeling obligation about how a specific asset was made.
A disclaimer on the page
These codes live in the image metadata and the product feed, where a person reading your site will never see them.

Why it matters to you

Almost everything else written about AI and content is judgement, which is hard to act on. This is a rule. A standards body maintains the vocabulary, and it applies when an image is made rather than when it is published. It also shows the direction of travel. The first concrete labeling duty has landed in product listings, and the machinery for recording how something was made now exists for anyone to use.

What to ask or check

  1. 01Does whoever makes our product images know which of the three codes applies to each one?
  2. 02Are AI-generated titles and descriptions being labeled as such in the feed, separately from the images?
  3. 03If a photograph was only edited with generative fill, is it being labeled as created rather than edited?

What people get wrong

That any image an AI tool touched is labeled the same way. IPTC separates media created by a generative model from media merely edited with one, and from ordinary retouching by a person using non-generative tools.

Red flags

  • Product images labeled as AI-created when generative fill only changed a background.
  • AI-written titles and descriptions going into a feed with no separate labeling.
  • A photographer or agency who has never heard of the metadata field the rule names.

Who owns it

Whoever produces the asset, in practice, because the code is written into the file then. Google sets the requirement for listings and IPTC maintains the vocabulary it points at.

Where you will see it

In product feeds and image files, set at the moment the asset is made rather than when the page goes live.

Scaled content abuse

Scaled content abuse is Google's name for generating many pages mainly to manipulate rankings rather than help people. Its definition says this applies no matter how the content is created, so using AI is not itself the violation. Google states it focuses on the quality of content rather than how it was produced.

AI Overviews

An AI Overview is the summary Google sometimes places above the results, with links to sources. Google says they are only shown when its systems judge them additive to classic Search, so they often do not trigger. There is no special file or markup that gets you into them.

Google-Extended

Google-Extended is a robots.txt token that controls whether content Google crawls from your site may be used to train Gemini models and ground Gemini apps. Google states it does not affect your inclusion in Search or your ranking. It is a training control, not a way to stay out of search results.

AI performance report

An AI performance report shows how often AI search features show or cite your site. Google's version counts impressions in AI Overviews and AI Mode. Bing's counts citations and states its numbers do not indicate ranking, authority or placement. Neither publishes a position inside an AI answer.

Grounding

Grounding is tying an AI answer to retrieved sources rather than letting the model answer from memory. Google uses the word as a synonym for retrieval augmented generation. Microsoft treats groundedness as something you measure on a scale, which is the more useful way to think about it.

Query fan-out

Query fan-out is the model turning one question into several searches at once. Google describes it as a set of concurrent, related queries generated by the model to fetch additional relevant search results. The trap is the obvious response: Google names a page per fan-out query as scaled content abuse.

Want this explained against your own numbers?

Twenty minutes, a straight answer, and no follow-up sequence if you decide not to work with us.