Chinese regulator drafts rules to protect teenagers from ‘intimate’ AI companions

China's Cyberspace Administration has introduced draft regulations to safeguard teenagers from potential harm caused by artificial intelligence-powered companions, including a ban on virtual intimacy services for individuals under 18. These rules, titled "Ensuring minors' safe and healthy use of the internet," aim to prevent the creation of virtual relatives or companions and prohibit services that could lead to addiction or negatively impact the mental and physical health of minors.

Additionally, the regulations would prohibit online social networking services among strangers, with the exception of those aged 16 and above who are self-supporting.

Platforms must activate minors' mode when operating online games, social sites, and AI services that could affect the cognitive development of users under 16. Operators are required to accurately identify underage users and prevent them from using algorithms that could induce emotional dependence, addiction, or excessive spending. Violations could result in fines up to 10 times the platform's income if it exceeds 1 million yuan (US$149,000), with confiscated income as punishment.

The proposals are currently open for public comment until October 17 and represent the latest restriction on screen time for teenagers. According to data from the China Internet Network Information Centre, individuals aged 6 to 19 accounted for 18% of all Chinese internet users, or 112.5 million people, as of the end of 2025. A lawyer and director at Geason Law Offices, He Qianggao, noted that defining a 'virtual intimate relationship' depends on whether a service simulates or substitutes real family or partner relationships and builds emotional bonds through character settings and emotional guidance.

He added that ordinary AI chats or functional role-playing would generally not fall under scrutiny, but services that encourage emotional dependence or replace real-world relationships could still be subject to regulation, even if not explicitly labeled as AI companions.

The compliance challenge for AI companies lies not only in verifying users' ages but also in determining which services they can access once identified as minors. This could lead to increased costs across product design, technical infrastructure, and compliance reviews. The latest regulations follow two months after China introduced its first dedicated rules governing humanlike AI interaction in July, targeting AI systems that excessively cater to users or encourage emotional dependence in a manner that damages real-world relationships.

In response, tech giants such as ByteDance's Doubao and Alibaba Group Holding's Qwen disabled customised agent functions that allowed users to create characters with fixed names, personalities, and speaking styles. Alibaba, the company behind Qwen, owns the South China Morning Post.