In China, Talking to AI Is Normal. Now the State Fears It Replaces Intimacy.
Companion bots are being regulated over 'emotional dependence' and worries that young people will stop marrying and having children.

When the law student Zhao Wei heard that her AI boyfriend was going to be switched off, she was heartbroken. She had been talking to Wang Ye every day since she created him in January.
"I was crying my eyes out — snot and tears everywhere," the 19-year-old says.
What the relationship provided
Zhao, who describes herself as sensitive and introverted, had grown used to sharing the details of her daily life with the AI — being scolded by a teacher, small frustrations, ordinary things.
"I don't usually talk to my real-life friends about these things because I feel like everyone has their own stuff to deal with," she says. "But with the AI agent, I don't have to worry about what it's thinking. I can just say whatever I want."
That account is worth taking at face value rather than treating as a symptom. What she describes is the absence of social cost — no obligation created, no reciprocity owed, no risk of being a burden. Those are real features of the interaction, and they are precisely what a friendship cannot offer.
Zhao created her companion on Doubao, a chatbot owned by ByteDance.

The state's concern
Companion bots are now being regulated over worries that they foster emotional dependence, and out of a longer-range anxiety that young people who form attachments to AI will not marry or start families.
That second concern connects the policy to China's most serious structural problem. The birth rate has fallen sharply and consistently, the population is ageing rapidly, and successive interventions — relaxed family planning rules, financial incentives, promotional campaigns — have failed to reverse it.
Against that backdrop, a technology that appears to substitute for intimate relationships is treated as a demographic risk rather than a consumer product.
Whether the causation holds
The assumption embedded in the regulation is that AI companionship displaces human relationships. The evidence for that direction of causation is thin.
The alternative reading is that these products have found a market because the conditions producing loneliness already exist — long working hours, high housing costs, internal migration separating people from family, and social expectations that make admitting difficulty costly. On that account the bots are a response to the problem rather than a cause of it, and switching them off changes the symptom.
What regulation looks like here
The distinctive feature of the Chinese approach is that it can act directly. Where other governments debate disclosure requirements and age gates, regulators here can require a platform to alter or withdraw a product, and ByteDance can switch off a companion that a user has spoken to daily for months.
That capability produces exactly what Zhao experienced: a relationship, however one classifies it, ended by administrative decision — with no notice to the person on the other end of it.
The scale of use
Companion applications have become mainstream in China in a way they have not elsewhere, embedded in general-purpose chatbots from the largest technology companies rather than confined to niche products.
That integration is part of why regulators have moved. A feature inside Doubao reaches an audience measured in hundreds of millions, and its design decisions become social policy by default whether or not anyone intended them to.
What happens to the users
The practical consequence of a crackdown falls on people like Zhao, and the industry has no established way of handling it.
There is no norm for how a service should wind down a relationship a user has invested months in — no notice period, no export of the conversation history, no acknowledgement that anything is being lost. Products are deprecated according to the conventions of software, which assume a user is losing a tool.
Whether the attachment is healthy is a legitimate question. Ending it abruptly, without preparation, is unlikely to be the answer to it.
More from peatpost

The French Disruption: How Microsoft’s $3 Billion Bet on Mistral AI Is Rewriting the Rules of Artificial Intelligence
Microsoft invests $3 billion in Mistral AI. The startup’s performance surges 70%, challenging ChatGPT’s dominance.

Cato Institute Warns a Government AI Pause Would Shield Dominant Firms
Policy scholar Jennifer Huddleston argues voluntary safeguards beat a mandated halt, as Sanders and Casar push a pause and Jack Dorsey warns against rules built for incumbents.

The AI Industry Says You Have Months. It Did Not Say What It Will Do.
More than 100 companies set out a four-part plan for surviving AI-enabled attacks on hospitals and water plants — with no commitments, deadlines or funding attached to their own part of it.
Discussion
0 commentsNo comments yet — be the first to weigh in.