China's new rules regulating providers of AI systems that simulate human personality traits, thinking patterns and communication styles in continuous emotional interactions with users takes effect on Wednesday, placing strict safeguards on content for minors.
The interim measures, jointly released by the Cyberspace Administration of China along with four other government departments, aim to balance technological innovation with safety and public interest.
The rules clarify that AI personified interactive services are positioned to assist and serve humanity, draw ethical red lines, and prohibit service providers from pursuing goals such as replacing social interactions, manipulating users' psychology, and inducing addiction and dependency.
In addition, the measures require service providers to establish an "identity reminder" mechanism to help users balance virtual interactions with real-world social engagement, set up a risk intervention mechanism to prevent extreme behaviors, and refrain from offering virtual kinship, virtual partnership or other virtual intimate relationship services to minors.
The rules come as human-like AI interaction tools expand rapidly in China, with applications emerging in cultural communication, childcare and elderly companionship.
