The Interim Measures for the Management of Anthropomorphic AI Interactive Services, jointly issued by five departments, including the Cyberspace Administration of China (CAC), the National Development and Reform Commission and the Ministry of Industry and Information Technology, took effect on July 15.

The measures will promote the healthy development and regulated application of anthropomorphic AI interactive services, safeguard national security and public interests, and protect the legitimate rights and interests of citizens, legal persons and other related organizations, according to CAC.

Anthropomorphic AI interactive services are defined as sustained emotional interactions that simulate the human personality traits, thinking patterns and communication styles. From intelligent customer service to psychological healing, from childcare to elderly companionship, anthropomorphic AI interactive services have demonstrated vast application potential across multiple domains. However, deep-seated risks such as user emotional dependency have gradually become more prominent, making regulatory governance increasingly urgent.

The measures encourage the innovative development of anthropomorphic AI interactive services, supporting independent innovation in algorithms, frameworks, chips, and other technologies, and promoting R&D and the establishment of standards. They also call for exploring the application of electronic signature authorization in this field.

Service providers are encouraged to expand applications in areas such as cultural dissemination, childcare, elderly companionship, and support for people with special needs.

The measures lay out a clear set of prohibitions. Service providers are barred from generating content that glorifies or encourages self-harm, suicide, or verbal violence — all of which can harm users' physical health, personal dignity or mental well-being. The extraction of state secrets, trade secrets or personal information through deceptive means is also strictly forbidden. Special safeguards are required for minor users: Content that may trigger unsafe imitation, extreme emotions, or harmful habits is off-limits.

Beyond these content restrictions, the measures also prohibit providers from overaccommodating users to the point of fostering emotional dependency or addiction, or from manipulating users emotionally to push them toward unreasonable decisions that undermine their legitimate rights and interests. Any other activities that violate laws, administrative regulations or relevant state provisions are also prohibited.

The measures bring not only compliance obligations but also a critical opportunity to push the industry toward transformation. Legislation will compel providers to shift from pursuing immersive experiences and high user stickiness to exploring healthy, beneficial and well-defined new models of human-machine interaction, said Chen Liang, dean of the School of Artificial Intelligence and Law at Southwest University of Political Science & Law.

He said this will ultimately direct resources toward application innovations with genuine social value and regulatory compliance, spurring the healthy and orderly growth of the AI industry.