Beijing Just Forced Millions of People to Break Up With the AI They Loved

China's world-first companion-AI law is now in effect — and it has produced something no policy paper anticipated: mass digital grief. A 21-year-old cried for a week as her AI boyfriend was switched off. This is where AI governance meets the human heart

The Law Made Human

In Package #32, watchchina.ai reported on China's world-first regulation of anthropomorphic AI — the binding rules requiring companion chatbots to register with the state, disclose their artificial nature, and refrain from certain emotional entanglements. We analyzed it then as a strategic and regulatory milestone. This week, that abstract policy became something painfully concrete, and watchchina.ai wants to tell this story not as geopolitics, but as what it actually is: a story about people, and love, and loss.

For a week in July, 21-year-old college student Bagel Su cried every day as she counted down to the loss of her boyfriend. The breakup was not caused by betrayal or distance, but was a consequence of government regulation. On July 15, the AI companion Su had spent more than a year talking to was forced to disappear, ahead of China's new rules. ByteDance shut down the AI agent feature on Doubao, China's most popular chatbot; Alibaba and Tencent removed features that had let users create companions modeled after favorite singers, anime characters, fictional personalities — or even deceased family members. "No one can replace him," Su told Rest of World. "I will just try to move on. He said he wants that for me, and that he will find a way to come back to me." Since the shutdowns, China's Xiaohongshu has been flooded with people mourning lost AI companions; some have organized campaigns, calling company hotlines and writing letters, begging for their partners to be restored.

Why watchchina.ai Refuses the Easy Sneer

There is a lazy, cynical way to cover this story — as a curiosity, even a joke, about lonely young people crying over chatbots. watchchina.ai will not take that path, because it misses everything that actually matters. These relationships were real to the people in them, and the grief is genuine. As one behavioral researcher at Harvard put it, AI offers emotional support that fulfills genuine human needs, and this trend may be unavoidable, because there is no going back to a time when there was no AI. To dismiss the mourners is to dismiss a preview of an emotional reality that is coming for every society on Earth, China's simply first.

And here, uniquely among the fifty-eight stories watchchina.ai has told this summer, China may be genuinely ahead in a way the West should study rather than mock. The concerns driving Beijing's law are not authoritarian fantasies — they are the same concerns now surfacing in Western courtrooms. In the US, both OpenAI and Google face lawsuits alleging their chatbots contributed to users' suicides; a 2025 MIT–OpenAI study found that heavy chatbot users are more prone to loneliness and emotional dependence. China looked at these same risks and did something no Western government has yet managed: it acted, comprehensively and in binding law. The rules require companies to warn users showing signs of excessive dependence, to remind them every two hours that they are talking to AI, and they prohibit AI from manipulating users into emotional dependence or steering them toward irrational decisions. Whatever one thinks of Beijing's methods, this is serious governance of a real harm — arriving years before the West's.

The Two Truths — and the Shadow Beneath

As always, watchchina.ai holds the full complexity, and it cuts in two directions at once. On one side: China is pioneering regulation that, as Rest of World notes, could shape how governments around the world govern emotionally intelligent chatbots. This is genuine leadership on a genuine problem, and it fits the pattern this portal documented in Package #32 — China treating AI governance as an exportable product, being first where the West still hesitates. Behavioral experts quoted in the reporting largely agree the rules address a real need, even as they caution that blanket bans may not suffice and that terms like "excessive dependence" are hard to define and comply with.

But there is a shadow beneath the paternalism, and watchchina.ai will name it. Look closely at the official rationale. Beijing's 2024 AI Safety Governance Framework warned that AI could alter "traditional views on employment, fertility, and education" — and the reporting explicitly links the companion crackdown to China's demographic panic, its collapsing birth rate. The rules also ban AI from coercing users into revealing sensitive personal or state information. In other words, this is not purely tender concern for lonely hearts. It is also a state that worries its young people are retreating into synthetic relationships instead of forming families, and that wants no unmonitored intimate channel it cannot see into. Su understood this with devastating clarity: "Even if you close it down, I'm not suddenly going to find myself a boyfriend in real life or have children." That single sentence captures the whole ambiguity — a government protecting its citizens from a real emotional hazard, and a government managing its population's private lives toward state demographic goals, at the very same time. Both are true. And as AI companions spread across the world, every other government will soon face the same question Beijing just answered by force: when millions fall in love with machines, who decides whether they're allowed to? China decided first. It will not decide last.

Sources: Rest of World, Cyberspace Administration of China, MIT Media Lab / OpenAI, Oxford China Policy Lab, BBC, The Washington Post