Content policy and blocked prompts

What is blocked

Studio does not generate sexual content involving minors, other illegal content, sexually explicit imagery, hateful or harassing depictions of real groups, content that infringes someone else's intellectual property, or realistic imagery of identifiable real people without their consent. These limits apply to both pack and Canvas mode.

How the safety classifier behaves

Prompts and generated images are screened by a safety classifier. A request that the classifier judges to be disallowed is blocked before any image reaches you, and the blocked request does not count against your allowance. Borderline requests can sometimes be blocked even when your intent was allowed; rewording the prompt usually resolves this.

Appeals and overrides

If you believe a prompt was blocked in error, you can report it from the blocked-result notice and a moderator will review it. Repeated attempts to produce disallowed content can lead to your access being limited. For how references are screened, see the Canvas mode article.