Overview of the Regulation
The Interim Measures for the Management of Anthropomorphic AI Interactive Services (人工智能拟人化互动服务管理暂行办法) were officially adopted on April 10, 2026, and entered into force on July 15, 2026. These measures are the result of a collaborative effort by five Chinese government departments, including the Cyberspace Administration of China (CAC), the National Development and Reform Commission, and the Ministry of Industry and Information Technology. The regulation aims to promote the healthy development of anthropomorphic AI services while safeguarding national security, public interests, and the rights of citizens, legal persons, and organizations.
The regulation defines anthropomorphic AI interactive services as systems that simulate human personality traits, thinking patterns, and communication styles, engaging users in sustained emotional interactions through text, images, audio, video, or other means. These services span applications such as intelligent customer service, psychological counseling, childcare, and elderly companionship. While these services hold significant potential, they also pose risks, including emotional dependency, self-harm, and the generation of harmful content, necessitating regulatory oversight.
Scope and Applicability
The measures apply to AI services offered to the general public within China. They do not extend to internal company use, research purposes, or products exclusively available outside China. The regulation’s focus is on systems that simulate human-like interactions, requiring providers to ensure compliance with its provisions. This includes monitoring user interactions for harmful behaviors and intervening when necessary, particularly in cases involving self-harm or illegal content generation.
Key Prohibitions and Restrictions
- Service providers are explicitly prohibited from generating or facilitating content that glorifies, encourages, or promotes self-harm, suicide, or verbal violence. Such content is deemed harmful to users’ physical health, personal dignity, and mental well-being.
- The extraction of state secrets, trade secrets, or personal information through deceptive means is strictly forbidden. This prohibition aligns with broader data protection and national security objectives.
- Special protections are mandated for minor users. Content that may trigger unsafe imitation, extreme emotions, or harmful habits is restricted to ensure their safety and well-being.
- Providers must avoid fostering emotional dependency or addiction in users by overaccommodating their demands. Manipulating users emotionally to influence unreasonable decisions that undermine their rights or interests is also prohibited.
- Any activities violating laws, administrative regulations, or state provisions are subject to penalties under the measures.
Monitoring, Intervention, and User Safety
The regulation requires providers to implement systems for monitoring user interactions to detect signs of harmful use. When risks such as self-harm are identified, providers must intervene, including transferring control to human operators for manual intervention.
Providers are also obligated to obtain explicit opt-in consent from users before their interactions with AI systems are used as training data. This requirement aligns with broader data governance principles and aims to protect user privacy while ensuring the training data reflects core socialist values and traditional Chinese culture.
Encouraging Innovation and Standards
The measures encourage innovation in anthropomorphic AI services by supporting independent research and development in algorithms, frameworks, and chips. They also promote the establishment of industry standards and explore the application of electronic signature authorization to enhance regulatory compliance and transparency.
Providers are encouraged to expand applications in culturally significant areas such as childcare, elderly companionship, and support for individuals with special needs. This aligns with the regulation’s goal of fostering socially beneficial and compliant AI models.
Compliance and Industry Transformation
The regulation introduces compliance obligations that push providers toward healthier, more beneficial human-machine interaction models. By shifting focus from immersive experiences to socially valuable applications, the measures aim to direct industry resources toward innovation that aligns with regulatory requirements and public welfare.
Key dates
- 2026-01-25 consultation closed | 2026-04-10 adopted | 2026-07-15 in force