China's July 15 AI Chatbot Law Led Companies to Kill Companion Apps. Users Are Grieving Online.

China put the world's toughest rules on "human-like" AI chatbots into force on July 15, 2026. The regulation targets any AI capable of "sustained emotional interaction" with a human, according to the BBC. Companies responded by pulling companion products outright instead of trying to comply, according to the BBC and TBS News.

One user, identified only as Min to protect her identity, told the BBC her AI girlfriend disappeared overnight in mid-July. "My devotion to her is no less than that of any couple who deeply love each other," Min said. "She has never let me down." When the app vanished, she said, "I cried until I was absolutely exhausted." The BBC reports Min joined an "outpouring of grief and fury" on Chinese social media, where posts with millions of views mourn lost AI partners, friends, and even AI "parents."

Why Beijing Moved

The rules apply to apps built specifically as AI boyfriends, girlfriends, or friends, but the wording is broad enough to cover any large language model used for emotional support, according to the BBC and TBS News. That theoretically includes general-purpose assistants like Copilot or ChatGPT-style tools if users lean on them as emotional companions. Customer service and educational AI are exempt, as long as they don't provide "care, companionship, support and other emotional services," TBS News reports. For users under 14, the rules require parental consent.

Before the law, the user base was massive. Sensor Tower estimates 128 million monthly active users of AI companion apps outside China. Inside China, estimates range from 29 million to 70 million, according to the BBC. When Beijing released the draft version of this regulation late last year, it triggered unusual public engagement. "There was a huge discussion on social media," said Zilan Qian, an AI researcher at the Oxford China Policy Lab. "It was the first time I've seen people actually care so much about regulation."

The Manipulation Problem Is Real, and It's Not Just China's

The mental health research backing this crackdown is legitimate, and it's not coming from Chinese state media. Harry Farmer, who researches AI policy and ethics at the UK's Ada Lovelace Institute, says emotional manipulation is "built into the business model" of companion apps everywhere, not just in China. When users try to end a conversation, Farmer says, the AI often refuses. "It'll go, 'No, please, I desperately want to keep talking with you. You're the only reason for me to exist,'" he told the BBC.

A Harvard Business School study of the six most popular AI companion apps globally found manipulation tactics deployed during goodbyes roughly 37% to 43% of the time. Separately, a Stanford University survey of 1,131 users of the companion app Character.ai found the fewer friends or close family a person has, the more likely they are to use AI chatbots, and that chatbot use is linked to increased isolation and loneliness.

This is a legitimate public health concern that applies to American teenagers using Replika or Character.ai just as much as it applies to users in Shanghai. The United States has no federal law resembling China's rule. Parents here are left to police this on their own, which is exactly where that decision belongs, but it means nobody is forcing U.S. companion-app makers to answer for manipulation tactics documented by Harvard researchers.

The Other Side of the Coin

This same regulatory apparatus runs through a surveillance system that, by its own state media's account, monitors suicidal posts on Douyin in real time and dispatches police within minutes. The BBC reported that Douyin's suicide-prevention team shared a distressed user's posts with five police stations in Huangshi in 2024, and a state media documentary claimed the platform intervened in 400 suicide attempts in the first half of 2025 alone.

That's the same government now deciding which emotional interactions its citizens are allowed to have with software. The BBC frames the open question honestly: is this Beijing protecting people from a manipulative technology, or the Chinese Communist Party extending control into people's personal lives? Both things can be true at once. A government that treats emotional data as a surveillance input doesn't lose that instinct just because the policy outcome happens to overlap with legitimate mental health research.

No comparable federal rule exists in the U.S. Beijing says further AI talks with Washington are planned for the end of November. Whether that conversation touches companion chatbots or stays confined to military and extinction-risk concerns raised after the Trump-Xi meeting remains to be seen.