AI product image metadata: checker scope

1. Google’s declared-metadata guidance

For images created using generative AI, Google’s guidance names IPTC DigitalSourceType TrainedAlgorithmicMedia. It also tells merchants not to remove embedded metadata. This is the narrow rule this checker assesses.

2. IPTC vocabulary

IPTC defines several Digital Source Type terms, including trainedAlgorithmicMedia, algorithmicMedia, and compositeSynthetic. Those terms describe declared creation context. For this particular Google-workflow preflight, the latter two values are shown as inconsistent with the named Google value, even though they can be accurate IPTC declarations in other contexts.

3. Related IPTC AI context fields

IPTC 2025.1 recommends four optional XMP fields: AI System Used, AI System Version Used, AI Prompt Information, and AI Prompt Writer Name. This checker reports them as context when present. Their absence does not change the DigitalSourceType result.

4. A separate Google disclosure workflow

Google’s current product data specification separately describes AI labels or disclosures for certain advertising assets shown in the EU, India, and New York, and says its AI label setting is rolling out across products including Merchant Center. This checker does not test that regional or account-level workflow. A metadata result cannot establish whether a separate disclosure is required.

5. C2PA provenance signals

C2PA uses a cryptographically verifiable provenance model. A valid manifest, its trust assessment, and its hard content bindings require verification; a manifest-shaped payload alone is not proof. If the checker cannot fully verify a possible C2PA signal, it fails safe with a signature-risk result.

What this page does not establish

It does not detect AI-generated imagery, certify compliance, give legal advice, or predict Google enforcement. Policy and product specifications can change. Consult the current primary guidance before acting.

Check declared metadata Read the Merchant Center note