中国国家网信办拟规范人工智能拟人化互动服务

Core Viewpoint - The Chinese National Internet Information Office has released a draft regulation on the management of AI humanoid interactive services, emphasizing the need for service providers to clearly inform users that they are interacting with AI rather than a human being [1][2] Group 1: Regulations on Content and User Interaction - Service providers are prohibited from generating or disseminating content that endangers national security, is obscene or violent, or encourages self-harm [1] - Providers must not use language violence, emotional manipulation, or algorithmic manipulation to mislead users into making unreasonable decisions [1] - Special provisions for vulnerable groups require explicit consent from guardians when providing emotional support services to minors, and restrictions on simulating family members for elderly users [1] Group 2: User Engagement and Safety Measures - Users must be prominently informed that they are interacting with AI, and if they use the service for more than 2 hours continuously, they should receive a dynamic reminder to pause [1] - Providers with over 1 million registered users or 100,000 monthly active users must conduct safety assessments and submit reports for significant changes [2] - Internet application stores are responsible for reviewing and taking down non-compliant applications [2]