Pre-Screening and Post-Screening Policy
This policy describes how Girlfriendly AI screens content both before it is created or published (pre-screening) and after it is live on the platform (post-screening). All content on Girlfriendly AI is AI-generated; screening applies to prompts, character definitions, generated media, and anything made publicly visible.
Pre-screening: before generation
Every generation request passes through automated screening before any content is produced:
- prompts are evaluated by automated classifiers, and requests indicating prohibited content — including any depiction of minors or other illegal material — are blocked before generation begins
- character definitions, names, ages, backstories, and appearance descriptions are screened when created or edited, not just at generation time
- generated outputs are screened again before they are delivered to the user
Pre-screening: before publication
Content that a user asks to make public is held to a stricter standard than private content. Every character and associated media submitted for public visibility is reviewed by a trained human reviewer before publication. Nothing appears in public areas of the platform without passing both automated screening and human pre-publication review.
Post-screening: after publication
Screening continues for as long as content remains on the platform:
- automated monitoring continuously re-evaluates published content against current classifiers
- content is re-reviewed by a person whenever it is reported through our Report Abuse / Takedown Policy page or support channels
- periodic sweeps re-examine existing public characters and media, including content published before a standards or classifier update
- content found to violate current standards is removed even if it passed earlier screening
Human oversight
Automated screening determines what is blocked instantly, but people supervise the system: moderators review flagged and borderline content, audit classifier decisions, and can override automation in either direction. Serious enforcement decisions — account termination, illegal-content determinations, and appeal outcomes — are always made by a person.
Consequences of failed screening
Content that fails screening is blocked or removed, and repeated attempts to bypass screening — through re-wording, obfuscation, or coded language — are treated as violations in their own right under the enforcement ladder in our Content Moderation Policy.
Related policies
Removal requests and appeals are handled under our Content Removal Policy. Our Underage Policy governs the absolute prohibition on any depiction of minors. Questions about this policy can be sent to [email protected].