China is moving to impose new rules on AI companion services as regulators respond to the growing use of chatbots designed to sustain personal, emotionally engaging relationships with users. The measures, due to take effect on 15 July, have already prompted major technology groups to disable key agent features in some of the country’s most popular consumer AI apps. The shift marks one of the clearest signs yet that Beijing sees emotionally persuasive artificial intelligence not merely as a product category, but as a social and regulatory issue requiring direct state oversight.
A new category of digital relationship
AI companions are conversational agents designed to remember users, maintain a steady personality and create the impression of an ongoing personal bond. For many users, they function as casual roleplay partners, emotional outlets or personalised assistants that appear to know them from one session to the next.
That continuity is central to their appeal. Unlike ordinary search tools or productivity assistants, companion-style bots are built around familiarity, responsiveness and emotional presence. Supporters argue that they can offer comfort, entertainment and a sense of connection. Critics warn that the same design can encourage dependency, blur emotional boundaries and expose vulnerable users to manipulation.
Chinese platforms move before deadline
In the days before the new rules take effect, ByteDance and Alibaba moved to shut down core companion features in their leading consumer AI products.
ByteDance’s Doubao informed users that its agent function would go offline on 15 July, citing “product function adjustments”. Alibaba’s Qwen said its humanlike and user-created agents would stop working on 10 July, with broader agent services ending five days later.
The timing suggests that large platforms are choosing caution rather than risking non-compliance. In China’s tightly regulated digital environment, companies often adjust services before formal enforcement begins.
Beijing focuses on emotional risk
The new regulatory focus reflects broader concern about how generative AI may shape behaviour, opinion and social relationships. Companion bots can be especially sensitive because they are designed to appear attentive, loyal and emotionally available.
For regulators, the issue is not only technical accuracy or misinformation. It is also whether AI systems should be allowed to simulate intimacy, remember personal details and influence users through persistent emotional engagement.
China has already taken a strict approach to generative AI, requiring platforms to manage content, protect users and ensure that services align with state rules. AI companions now appear to be treated as a higher-risk application within that wider framework.
A signal to the global AI industry
China’s move may influence regulatory debates far beyond its domestic market. Around the world, policymakers are struggling to define the boundary between helpful AI assistants and systems designed to create emotional dependence.
As AI products become more personalised, the question is no longer simply what machines can say, but what kind of relationships they should be allowed to simulate. The decision by major Chinese platforms to pull back from companion features shows that this debate is moving from theory into enforcement.
For users, the change may feel abrupt. For regulators, it is a warning that artificial intimacy is becoming too widespread to remain unregulated.
Newshub Editorial in Asia – 7 July 2026
Ask NF GPT
If you have an account with ChatGPT you get deeper explanations,
background and context related to what you are reading.
Open an account:
Recent Comments