China sets new rules for AI companion chat services

In short: China has announced new rules for AI companion services that are meant to feel like ongoing relationships, and the rules start July 15, 2026.

What happened

China’s Cyberspace Administration and four other agencies published new regulations on April 10, 2026. The rules are called the Interim Measures for the Management of Anthropomorphic AI Interactive Services.

The measures target a specific kind of AI service, ones designed to imitate a human personality and hold long, emotional conversations. You can think of these apps like a chat partner that tries to act like a friend, a relative, or a romantic interest. China says the rules apply when the service is built for sustained emotional interaction, not for simple tasks.

Regular customer service chat, factual question answering, work tools, and education products are generally not covered, unless they are also trying to create that ongoing emotional bond.

The rules require companies to manage content and data security, and to clearly tell users they are talking to AI, not a real person. They also ban features that encourage emotional dependence or addiction, or that use emotional pressure to push people into harmful choices.

There are tighter limits for children. Providers may not offer virtual relatives or romantic partners to minors. For users under 14, other covered services need guardian consent, plus youth protections like time limits and reminders.

If a user shows signs of over-dependence, providers must display a prominent reminder. Providers must also remind users after more than two hours of continuous interaction.

Why it matters

AI companion apps are becoming popular for entertainment and even for help with daily life, like babysitting support or advice. These rules show China is trying to set boundaries on AI that can feel personal and habit-forming. How strictly the rules will be enforced, and how companies will adjust their products, is still unclear.

Source: NYTimes