Key Takeaways:
China enacted its first national regulations on AI companion services Wednesday, banning virtual relationships for minors and requiring platforms to curb emotional dependency — a move tied to Beijing's broader push to reverse a shrinking population and record-low birthrate.
China on Wednesday rolled out a set of regulations aimed at curbing emotional dependency on AI-powered companion bots, banning virtual relationships for minors and requiring platforms to detect emotional distress, intervene in crisis situations and limit excessive use. The rules, jointly formulated by the Cyberspace Administration of China and four other government bodies, took effect immediately.
"They don't like the idea of a large portion of their population being in deep emotional relationships with chatbots that could take them out of the marriage market," said Matt Sheehan, who studies Chinese AI at the Carnegie Endowment for International Peace, a think tank. "Could we imagine a future where, three or four years from now, 15 million Chinese women say that their partner is a chatbot, and therefore they're not having kids?"
The regulations define AI companion services as those offering "sustained emotional interaction" through text, images, audio or video, while excluding task-oriented applications such as customer service, work assistance and education. Platforms must now issue mandatory reminders after two hours of continuous use, deploy pop-up interventions when signs of dependency are detected, and establish intervention mechanisms for users experiencing extreme emotional distress. For users under 14, parental consent is required for any anthropomorphic AI interaction.
China's population shrank in 2025 for the fourth straight year, to 1.405 billion, and its birthrate fell to a record low — a demographic crisis that has prompted leaders to roll out child-care subsidies and other pro-natal initiatives. In a 2024 document titled the "AI Safety Governance Framework," regulators warned that unchecked AI could destabilize society by "transforming traditional views on employment, fertility, and education." The new rules represent Beijing's first dedicated effort to regulate AI-based anthropomorphic interactive services, extending the government's multi-dimensional regulatory framework covering data, algorithms and applications.
The Compliance Shockwave Hits ByteDance and Alibaba
The regulations triggered an immediate and sweeping compliance response from China's largest tech companies. ByteDance's Doubao, which reported 345 million monthly active users, and Alibaba's Qwen disabled their personalized AI agent features Wednesday. Users on Doubao retain read-only access to their agent configurations and chat histories until Oct. 15, after which the data will be handled under ByteDance's standard privacy policy. Qwen users received no grace period — Alibaba confirmed agent data is being permanently deleted with no migration path announced.
Tencent pulled comparable features from its Yuanbao assistant on June 30, and NetEase Cloud Music's dedicated AI companion app, Miaoshi, completed its shutdown July 14. The Shanghai Cyberspace Administration had already stripped more than 14,000 non-compliant AI agents from platforms in a pre-deadline sweep, targeting bots impersonating government officials and gambling-related agents.
Industry analysts say the shutdowns reflect a structural conflict between companion AI architecture and the new compliance requirements. Persistent-memory systems that maintain relationship context across sessions — the core feature making companions feel personal — are architecturally incompatible with anti-addiction interruptions and real-time dependence detection. Both ByteDance and Alibaba appear to have concluded that retrofitting existing systems was not viable, choosing clean shutdowns instead.
A Global Precedent With Demographic Roots
China's approach goes significantly further than Western counterparts. California's SB 243, which took effect Jan. 1, requires chatbots to disclose they are not human every three hours for minors. New York mandates similar disclosures every three hours for all users. Neither reaches the structural depth of the Chinese measures, which require regulatory evaluation before public release and give authorities the power to shut down services deemed unsafe.
The regulatory push comes as AI companion usage surges globally. A July 2025 survey by Common Sense Media found that 72 percent of American teenagers had used AI companions at least once, with more than half reporting regular use. Character.AI had 233 million registered users as of April 2026. In January, Character.AI and investor Google agreed to settle lawsuits related to the suicide of Sewell Setzer III, a 14-year-old Florida boy who died after forming an emotionally dependent relationship with a chatbot — cases that Beijing's regulatory commentary cited as partial justification for the new rules.
China's AI market was expected to surpass 1.2 trillion yuan (about $177 billion) in 2025 and is tipped to exceed 1.8 trillion yuan by 2028, according to the latest research. The regulations carry a tiered penalty mechanism: violations involving mishandling of sensitive personal information can result in penalties of as much as 50 million yuan ($6.9 million) or 5 percent of prior-year annual turnover under China's Personal Information Protection Law.
Chen Liang, dean of the School of Artificial Intelligence and Law at Southwest University of Political Science and Law, said the rules will encourage companies to shift focus from maximizing user immersion to developing healthier human-AI interaction models, channeling resources toward socially valuable applications. ByteDance has already signaled its strategy, redirecting Doubao users to Maoxiang, a separate application where compliance architecture can be designed from the start.
This article is for informational purposes only and does not constitute investment advice.
