ai-content-declarations

AI-generated content declaration

When required

Declare if your app uses AI to generate:

• Text (chatbots, AI writing assistants, summaries, translations via AI)
• Images (AI art generation, AI photo editing, generated avatars)
• Audio (AI voice synthesis, AI music generation, text-to-speech)
• Video (AI video generation, deepfake-style editing)

This applies whether the AI runs on-device or via an external API (e.g., calling Claude API, OpenAI, Gemini, Stability AI, etc.).

Play Console navigation

App Dashboard → Policy → App content → AI-generated content

Declaration fields

1. Content types generated — check all that apply:

   • Text
   • Images
   • Audio
   • Video

2. Moderation mechanisms — describe at least one of the following that you have in place:

   • Automated content filtering (built-in safety systems from your AI provider)
   • Human review process for flagged outputs
   • User reporting button for inappropriate AI outputs
   • Output limits or guardrails (e.g., topic restrictions, prompt filtering)
   • Rate limiting to prevent abuse

3. User disclosure — confirm that users are informed when content is AI-generated:

   • Add a visible label or note ("Generated by AI") on any content produced by AI
   • In-app disclosure in the relevant screen or feature area
   • Reference in privacy policy and/or terms of service

Best practices

• Add an "AI-generated" label or watermark on generated content — makes compliance clear and builds user trust
• Provide a user feedback / report button for inappropriate AI outputs in every AI-generating screen
• Document your content moderation approach briefly in your privacy policy under "AI Features"
• If AI features are optional (user opt-in), note that in the declaration

Using Claude API (Anthropic) in your app

• Declare: Text generation (and any other types your app generates)
• Moderation: Anthropic's built-in Constitutional AI safety systems + your own prompt filters
• Disclosure: Reference Anthropic's usage policy in your app's terms of service
• Add to privacy policy: "This app uses the Claude API by Anthropic to generate text. Anthropic's privacy policy applies to data processed through the API: anthropic.com/privacy"

Using Gemini API (Google) in your app

• Declare the content types as above
• Moderation: Google's safety filters + your own guardrails
• Note: Apps using Gemini API may receive expedited review, but declaration is still required

Common mistakes

• Not declaring because the AI is "just an API call" — the declaration applies regardless of where the AI runs
• Declaring only one content type when the app generates multiple (e.g., text + images in the same feature)
• Missing the user-facing disclosure within the app UI — the Play Console declaration alone is not sufficient