Abstract
Prohibited content, prohibited activities, resource use rules, AI-specific restrictions (no autonomous high-risk actions without oversight, no undisclosed AI in customer-facing contexts, no biometric inference, no social scoring), channel rules, enforcement procedure.