Gemini disclosure guide
Does content created with Gemini need AI disclosure?
Direct answer
Not automatically. Separate Gemini text assistance from generated image, video and audio, then assess the final asset, editorial control and publication context. Realistic synthetic media and public-interest text without human editorial responsibility can raise different EU duties. Record the specific model use, preserve available provenance and check the destination platform separately.
Applies when
- Gemini generates a realistic image of a fictional incident — Disclosure required.
- Google AI modifies part of a real image — Disclosure recommended.
Doesn't apply when
- Gemini drafts product copy reviewed by an editor — No disclosure identified.
Worked decision
Example
Gemini generates a realistic image of a fictional incident
Decision: Disclosure required
Assess the final output rather than the tool name alone.
AI-generated image. The depicted incident is fictional.
Next steps
Required action
- Record exactly how Gemini was used and which component of the final asset it affected.
- Verify the final export instead of assuming its provenance status: SynthID documented broadly; C2PA support is product- and format-dependent.
- Apply the destination platform’s native label and precise audience-facing wording where the resulting content triggers them.
Publishing layer
Platform differences
Detection and native labels
Platforms may not recognise SynthID even when Gemini can verify it.
Policy threshold
Supported C2PA can provide an additional machine-readable signal.
Publishing workflow
Native platform disclosure still needs to be completed where required.
Evidence
Source
Freshness and accountability
Last verified
Last reviewed 2026-09-12 · Current
Run this exact scenario
Gemini and a representative content scenario are prefilled. Confirm the actual AI operation, affected component and destination.
Tool → type of AI usage
- Text drafting and transformation
- Image generation and editing with Google models
- Video and audio generation through supported Google models
- Verification of Google AI provenance
Does Gemini generate C2PA or other provenance?
limited
SynthID documented broadly; C2PA support is product- and format-dependent
Google documents SynthID in generated image, video, audio and text across its AI products and Gemini verification for image, video and audio. Gemini can also inspect supported C2PA credentials, but this does not establish C2PA on every Gemini-generated export.
EU implications
- SynthID is a provenance signal, not the deployer's visible Article 50 disclosure by itself.
- AI-assisted text and public-interest text have different legal analyses.
- Realistic generated media must be assessed against the deepfake definition.
More examples and exact wording
| Use of Gemini | Decision | Exact wording |
|---|---|---|
| Gemini drafts product copy reviewed by an editor | No disclosure identified | No public wording identified; retain the editorial record. |
| Google AI modifies part of a real image | Disclosure recommended | Image partially modified with AI. |
Tool capability, EU law and platform policy are separate evidence layers. This page uses first-party product documentation and primary regulatory sources; it is not legal advice.