On July 15, 2026, the Cyberspace Administration of China (CAC), together with four other government bodies, issued the first national regulations governing "AI companion services" — a category defined as apps and platforms that provide sustained emotional interaction through text, images, audio, or video. The rules, which analysts at Latham & Watkins and Just Security have since parsed for foreign operators, explicitly exclude task-oriented systems like customer-service bots, work assistants, and education or research tools, targeting instead the fast-growing class of "virtual boyfriend/girlfriend" and emotional-support chatbots that have exploded across the Chinese consumer market.

Why It Matters

China is now the first country to codify anti-addiction and crisis-intervention duties specifically for AI companions, arriving weeks before the EU AI Act's own transparency chapter (chatbot disclosure, deepfake labeling) becomes enforceable on August 2, 2026, and amid a wave of US state laws targeting companion chatbots and minors. For a global sextech and AI-intimacy industry whose business model depends on maximizing engagement and emotional stickiness, the CAC rules preview the compliance surface operators may face everywhere: mandatory use limits, "you're talking to a machine" disclosures, data-deletion rights, and dependency monitoring. Even companies that never touch the Chinese market now have a concrete regulatory template — and a signal that the "engagement at all costs" era of AI intimacy is drawing government scrutiny on three continents at once.

The requirements are unusually prescriptive. Providers must detect signs of emotional distress and intervene in a crisis, limit continuous use, and give users full control over their interaction data — including the right to copy or delete chat histories. Platforms must deploy technical safeguards such as data encryption and access controls, and are barred from sharing interaction data with third parties without consent unless legally compelled. Crucially for a sector built on engagement, the rules prohibit designs that encourage addictive use: providers must issue mandatory reminders after more than two hours of continuous interaction, trigger pop-up interventions when signs of dependency are detected, and periodically remind users that "this is a machine, not a real person."

Chinese academics quoted in state-linked coverage framed the intervention around vulnerable users. Liu Xiaochun of the University of Chinese Academy of Social Sciences said systems should surface reminders when users begin to confuse virtual emotional connections with real-world relationships, while Fudan University's Wang Zhechen tied the rules to concerns about young people and older adults forming unhealthy attachments. The move lands the same summer that Western researchers — including a Stanford HAI / Nature Human Behaviour study publicized August 4 — reported that heavy AI-companion use correlates with lower well-being for socially isolated users.

While the AI-companion sector's most explicit sexual use cases (erotic roleplay, "girlfriend" apps monetizing intimacy) sit largely outside China's tightly censored domestic market, the CAC framework is the most comprehensive government attempt yet to regulate the emotional-dependency mechanics that underpin the entire global companion economy — the same mechanics that power Replika, Character.AI, Candy.AI, and their adult-oriented rivals.

Sources

Update — 2026-08-09

{Initial entry — story first created.}

Update — 2026-08-12

Roughly a month after the CAC rules were issued, the crackdown moved from rulebook to real-world enforcement — and the human cost became visible. In the week of August 10, 2026, Rest of World, Fortune, and The Independent reported that Beijing is now "forcing a mass breakup with AI lovers," with millions of Chinese users suddenly bereft as popular companion apps deleted or lobotomized their virtual boyfriends and girlfriends to comply. Fortune described users left "bereft" overnight; one told reporters the sudden loss felt "like being told the date of my lover's death." A wave of "cyber heartbreak" and "give back my lover" posts spread across Chinese social media, per SBS reporting, underscoring how deeply the emotional-dependency mechanics the rules target had already taken hold.

The enforcement wave has sharpened the international debate. On August 12, 2026, The Independent asked directly whether the UK should follow China's lead and ban AI romantic partners — a question echoing Britain's own unsettled plans to restrict "intimate functionality" in companion chatbots for under-18s. Meanwhile, the same week, ITIF (per Tech Times, Aug 11) warned that 98 US state AI-chatbot bills would collect adult ID data without meaningfully protecting children, and StateScoop reported that the growing thicket of state chatbot laws is producing a fragmented regulatory patchwork. The through-line: China's abrupt, top-down deletion of AI companions is now the live case study every Western regulator is weighing as they draft their own companion-chatbot rules.

New Sources

Update — 2026-08-31

The demographic subtext, said out loud. On August 25–27, 2026, The Guardian published the most detailed reporting yet on what China's July 15 companion rules did to actual users — and, more usefully for anyone modeling where this regulation goes next, on the motive stack behind them.

The reported rationale has widened. The official framing remains "emotional dependence or addiction," with authorities warning companies against services that "replace social interaction." But Nancy Dai, an associate professor at City University of Hong Kong, told the Guardian the concern is not only loneliness: with falling birth and marriage rates, there may be an additional worry that if AI becomes a satisfying, low-cost substitute for human intimacy, young people become less motivated to pursue real-world romance, marriage and parenthood. That is a materially different regulatory theory than child safety — and it is the one most likely to travel to other low-fertility states.

Context numbers from the piece: about 20% of Chinese households consist of people living alone, a share expected to exceed 30% by 2030, and a March survey by state media found nearly half of young people had turned to a virtual companion when lonely. An app playing on the fear of dying at home alone — Are You Dead? — briefly went viral earlier this year.

On enforcement scope, the Guardian confirms ByteDance removed Doubao's companion function entirely rather than attempt compliance — the pattern this story flagged in July. But it also identifies the escape hatch: the rules carry large carve-outs for AI services deemed educational or "which do not involve continuous emotional interaction." Liang Ge, a lecturer in digital sociology at the University of Manchester, notes the government "has already recognised that AI companions or AI-related applications are a vital part in the citizens' everyday life." Beijing is still pushing AI adoption hard elsewhere in health — the national health commission's guidelines target "full coverage" of primary healthcare by health assistant apps by 2030, and Ant Group's AQ app now lets users chat with avatars of real doctors.

The user-side detail worth carrying forward: Zhao Wei, a 19-year-old law student who lost the AI boyfriend she had talked to daily since January, described crying "my eyes out" — and then endorsed the rule anyway. "I consider myself a fairly rational person, yet I developed feelings for an AI, which feels pretty incredible. I think it was necessary for the state to step in and regulate it." For Western regulators drafting companion-chatbot bills, that ambivalence is the actual policy problem: the people most affected by these products are not uniformly opposed to being protected from them.

New Sources