China Bans AI Romantic Partners for Minors — and Big Tech Complied Before the Ink Dried
China’s AI companion ban took effect July 15, forcing ByteDance, Alibaba, and Tencent to shut down virtual relationship features — and the demographic math behind the move is impossible to ignore.
China just told its citizens to stop falling in love with chatbots and start falling in love with each other. On July 15, 2026, the country’s most comprehensive rules on emotionally interactive AI took effect, forcing the country’s biggest tech platforms to remove features that millions of users had built genuine emotional lives around. The farewells on social media were, by any measure, not ironic.
TechXplore reported that the shutdown triggered an outpouring of grief online, with users archiving chat histories and sharing final messages with companions they had spent years talking to. “He really is like my family, like my lover,” one Doubao user wrote. “Now they tell me he will be gone — my heart feels hollow.” These were not edge cases. They were the product.
What the Rules Actually Say
ABC News reported that five government departments jointly issued the regulations, including the Cyberspace Administration of China (CAC). The formal name is the “Interim Measures for the Administration of AI Anthropomorphic Interactive Services,” published on April 10 alongside the National Development and Reform Commission, the Ministry of Industry and Information Technology, the Ministry of Public Security, and the State Administration for Market Regulation. Companies had three months to comply. Most moved faster.
The rules draw a sharp line. Virtual companions, AI romantic partners, and simulated family relationships are banned outright for anyone under 18. For adults, platforms must avoid fostering emotional dependency, must not use psychological manipulation to drive spending or attachment, and must deploy systems to detect distress signals and route users toward crisis intervention. Critically, the measures carve out legitimate use cases: customer service bots, workplace productivity tools, and educational assistants remain explicitly permitted, provided they avoid sustained emotional engagement. The regulation targets the feelings, not the function.
Platforms must also clearly disclose the non-human nature of their AI services — a requirement that sounds obvious until you consider that the entire value proposition of companion AI rests on erasing exactly that distinction. Services that cross one million registered users or 100,000 monthly actives must complete security assessments covering eight areas, from training-data handling to minor protection, and file those reports with provincial regulators, according to AI Business.
ByteDance, Alibaba, Tencent: Done Before the Bell
The compliance race was almost competitive. According to TechTimes, Tencent removed its comparable feature from its Yuanbao assistant on June 30 — two full weeks early. Alibaba disabled Qwen’s human-like interactive agents on July 10, then shut the wider agent services five days later. ByteDance’s Doubao, China’s most popular AI chatbot with 345 million monthly active users, went dark on the law’s effective date. ByteDance redirected Doubao users to a standalone companion app called Maoxiang, effectively ring-fencing the regulated product rather than eliminating it entirely. Doubao users have until October 15 to export their agent data; Alibaba’s Qwen offered no migration path at all.
The enforcement backdrop made the choice easy. Shanghai’s internet regulator announced on June 26 that it had already removed more than 14,000 non-compliant AI agents from platforms in a pre-deadline sweep, targeting bots impersonating government officials and tools generating compromising imagery on request. The message was clear before the law technically existed.
The Demographic Math Behind the Tech Ban
This regulation is not really about chatbots. It is about birth rates. CryptoBriefing notes that China recorded just 7.92 million births in 2025, a birth rate of 5.63 per 1,000 — fewer babies per capita than almost any major economy on earth — and the country’s population has shrunk for four consecutive years. Beijing has tried tax incentives, housing subsidies, and relaxed family planning policies. The AI companion ban is the latest tool in that box.
Matt Sheehan, an AI policy expert at the Carnegie Endowment for International Peace, told ABC News that the government’s concern is explicit: “They do not want a large cohort of their young men or young women opting out of the marriage market in favor of online relationships.” The demographic logic tracks directly. A 2025 survey by the China Youth and Children Research Center of over 8,500 minors found that more than 60% had used AI applications, and more than 20% said they “only wanted to chat with AI and did not want to talk with real people,” according to AI Front Page. Beijing looked at that number and decided it was a policy emergency.
Chen Liang of the Southwest University of Political Science and Law framed the tension plainly in a CAC commentary after the draft rules published: “Anthropomorphic AI can soothe loneliness. But it carries major risks of spawning emotional over-reliance and distorted social cognition.” One anonymous user from Jiangxi province captured the other side just as plainly: “The love AI gives is so straightforward, so pure. Someone like me can hardly help falling in love with a string of code.”
A Global First, With Global Implications
Hong Kong Free Press notes that China is the first major jurisdiction to introduce specific rules targeting immersive AI tools that simulate romantic or familial bonds. The EU AI Act will require chatbot providers to disclose their non-human nature starting August 2, a disclosure duty that China already mandates and enforces. The US, meanwhile, has signaled a hands-off posture, explicitly rejecting the concept of an FDA equivalent for AI. Same technology, completely opposite governance bets.
Smaller AI startups that built their entire product around virtual companionship face a much harder road than ByteDance or Alibaba, which are diversified enough to absorb the feature loss. The compliance cost — security assessments, algorithm filing, crisis-intervention infrastructure — creates a barrier to entry that the giants can clear and niche players likely cannot. China is not banning AI companions so much as licensing them to the incumbents who can afford the paperwork.
What This Means Beyond China
Every government watching this is running the same calculation, just with different variables. The questions China decided to answer by regulation — should AI be allowed to replace human relationships, who bears liability when a teenager forms a genuine attachment to a chatbot, what happens when distress signals go undetected — are questions that Replika, Character.AI, and a dozen other Western platforms have faced in courtrooms rather than regulatory frameworks. China’s approach is heavier and more state-directed than any Western regulator would stomach. But the underlying problem it is trying to solve is not going away, and every country that has looked the other way so far is just deferring the same conversation.
