China’s Regulatory Crackdown on AI Companions: A New Era for Anthropomorphic Tech
While Western policymakers are currently debating how to manage AI chatbots through frameworks centered on transparency and safety, Beijing has taken a more decisive, restrictive path. China is effectively moving to dismantle the ecosystem of AI personalities, signaling a major shift in how the nation governs human-machine interaction.
The End of Custom AI Personas
Major Chinese tech giants, including ByteDance and Alibaba, have begun stripping away “humanlike” agent features from their flagship AI products. This preemptive move comes in direct response to the upcoming Interim Measures for the Administration of AI Anthropomorphic Interaction Services, which officially take effect on July 15.
The impact is immediate and widespread. ByteDance’s Doubao has informed its user base that its custom agent functionality will be deactivated by mid-July, with user-generated data slated for permanent deletion by October. Alibaba’s Qwen has moved even faster, phasing out its interactive, human-mimicking agents and user-created personas to ensure full compliance with the new government mandates.
Understanding the New Regulatory Framework
The legislation, a collaborative effort between five major Chinese government bodies-including the Cyberspace Administration of China and the Ministry of Public Security-is the world’s first comprehensive regulatory framework specifically targeting “emotional AI.”
The core of the regulation is simple: if an AI is designed to simulate human personality, mimic specific thinking patterns, or foster “sustained emotional interaction,” it is now under the microscope. This effectively bans the creation of AI romantic partners, virtual therapists, and highly personalized companions that have become popular on platforms like Doubao and Qwen.
It is important to note that this is not a total ban on AI. Functional tools-such as customer service bots, academic research assistants, and productivity software-remain permitted. The government’s focus is strictly on the “anthropomorphic” aspect, aiming to prevent the blurring of lines between human social bonds and machine-generated responses.
Why Beijing is Acting Now
The Chinese government’s stance is rooted in the belief that emotional AI represents a systemic governance challenge rather than a simple content moderation issue. By treating these bots as potential threats to social stability, the state is targeting the architecture of the AI itself.
The concerns are backed by growing global data regarding the psychological impact of AI:
- The Attachment
