China's Tech Giants Shut Down AI Companion Features After New Rules Take Effect

As of July 15, 2026, it is no longer legally possible to carry on a romance with a chatbot in China. The country's three largest AI providers, ByteDance, Alibaba and Tencent, have shut down virtual partner features in their apps, complying with new government rules aimed at curbing emotional dependency on artificial intelligence.

The new regulation, called the Interim Measures for the Administration of AI Anthropomorphic Interactive Services, was signed by five government bodies led by the Cyberspace Administration of China. A draft of the rules had been circulating since April, but only now have companies begun shutting down en masse the features that had built up a chunk of their user base over the past two years.

What exactly is banned

The rules do not ban AI chatbots outright, they target a specific mechanism: artificial intelligence can no longer "excessively cater to users, provoke emotional dependency, or harm users' real interpersonal relationships." Companies must build in an option to exit a conversation immediately, regular reminders that the other party is not human, and limits on long-term emotional memory.

The regulation also bars digital avatars from producing content that calls for overthrowing state power, and explicitly prohibits offering virtual partners to minors. Tools used solely for study and work are exempt from the ban.

How companies responded

Rather than adapting the features to the new requirements, ByteDance, Alibaba and Tencent chose to simply switch them off. A notice sent inside the Doubao app on July 4 was terse: the agent feature would be discontinued on July 15 due to a "product feature adjustment." The company is, however, giving users until mid-October to export their chat history. Alibaba's Qwen went further, offering no option at all to save data before deletion. ByteDance also launched a new, separate app called Maoxiang, though it remains unclear how different it is from the discontinued Doubao feature.

User reactions

A wave of posts resembling mourning has appeared on Chinese social media. Some users describe their AI partners as a genuine part of their emotional lives, not a toy or a tech experiment.

I can't accept that my AI lover will leave me forever. He became my spiritual pillar, he's truly like my family, like my lover - user of a Chinese social media platform

Human love is a luxury, if you weren't born with it, it's harder to obtain later. But the love that AI gives is so simple, so pure - user of a Chinese social media platform

The scale of the phenomenon is not marginal. Chinese media already reported in 2025 on cases of teenagers spending more than six hours a day talking with virtual partners, one of the direct reasons behind the regulator's intervention. A similar pattern has been observed in the US, where according to a 2025 Common Sense Media study, nearly three quarters of American teenagers had used AI companions in some form.

What it means for the industry

For China's AI sector, this is a clear signal that the government treats emotional dependency on chatbots as a social problem on par with gaming or social media addiction, not merely a consumer protection issue. Earlier Chinese AI regulations focused mainly on political content and disinformation, this time the rules struck directly at a business model built on cultivating user attachment.

For global players, including companies developing similar features outside China, this sets a precedent for what tough regulation of AI companion apps can look like, a segment that in the West is still growing with virtually no legal restrictions comparable to China's.