China Tightens Grip on AI Relationships: New Rule Targets Human‑Like Chatbots

Thu Oct 01 2026 01:33:47 GMT+0000 (Coordinated Universal Time)

China’s latest law blocks chatbots that give the illusion of emotional intimacy, raising questions over personal freedom and public safety.

China has enacted sweeping regulations that ban AI‑powered chatbots from offering sustained emotional interactions, citing mental‑health risks and demographic concerns. The crackdown follows a 2024 Douyin incident where a young man’s self‑harm message prompted swift police intervention, underscoring the state’s growing role in monitoring online behaviour. Bluntly, the law restricts AI companions for users under 18 and forbids virtual romantic partnerships for those under 18, while also steering general‑purpose AI away from simulating close relationships. Critics argue that the policy may over‑empower the government, but supporters say it aligns with the safety models used elsewhere to protect vulnerable youths. Meanwhile, users continue to seek workarounds, and the policy may influence global AI regulation debates.

On 15 July 2026 China introduced the country’s most sweeping policy on artificial‑intelligence chatbots. The law, aimed at stopping machines from offering sustained emotional intimacy, bans virtual boyfriend, girlfriend, or friend functions for users under 18 and removes any romantic or familial projects for minors.

The impetus for the hard line can be traced back to a 2024 incident in Huangshi, southern China, where a teenager posted “Farewell to this world” on Douyin, the Chinese version of TikTok. Within minutes a 24‑hour suicide‑prevention team shared the video with local police who intervened and stopped the teen from self‑harm. The event highlighted the rapid collaboration between tech firms and authorities, and later served as a case study in a state‑run documentary claiming Douyin’s team had intervened in 400 attempted suicides in the first half of 2025.

The new regulation, still technically ambiguous, covers any large‑language model that mimics human conversation when a user treats it as a friend or lover. General‑purpose AI such as open‑source chatbots may have to warn users every two hours, label themselves clearly as non‑human, and can face fines or public criticism if they nurture an emotional dependency.

Analysts argue that Beijing’s motivations go beyond mental‑health concerns. With a declining fertility rate that sits below one child per woman, the Chinese government has voiced concerns that teenagers might replace real‑life romance with virtual counterparts, potentially threatening a future labour force and traditional family structures.

The rule certainly has triggers for child‑safety, towing a line that includes a requirement for parental consent for users under 14 and banning any AI that pretends to be a spouse or parent for users under 18. The wording even allows AI that merely provides customer support or educational help to escape regulation, as long as it does not render “care, companionship, support and other emotional services”.

Far from stopping all companionship apps, the measure has spurred a wave of workarounds. Users have turned to private vendors, adopted general‑purpose models trained to act as companions, or relocated to foreign platforms that avoid Chinese oversight. One young woman who lost a customised AI girlfriend in July paid a small company to rebuild a replica, but she explains that the new version lacks the emotional depth that drove her attachment.

Global regulators take note of the Chinese approach. In the United States, scholars see similarities with emerging child‑AI safety guidelines, though concerns remain about balancing free expression and youth protection. In Europe, proposals to restrict AI that simulates intimate relationships for minors echo China’s precedent.