Google Play AI-generated content policy: fixes and checks

The model may be supplied by another company, but generated content is still part of your app. Google’s policy requires compliance with its other content rules and an in-app way to report or flag offensive output without leaving the app.

Check generation and reporting together

Review input controls, output filters, and how users report a result. We’d follow a test report through to the developer’s handling process, rather than stop at a flag icon. Reports should inform filtering and moderation improvements.

Address the unsafe output

Remove features that generate prohibited content and repair the relevant safeguards. Include deceptive outputs and child-safety concerns in the assessment. For shared outputs, also review UGC requirements.

Test with safe representative prompts and benign report samples. Verify that refusal and filtering work in the app, not only in a provider’s demo. Give Google the corrected feature, report route, and moderation process. Avoid claiming a model can never produce a harmful answer.