China tells AI boyfriends and girlfriends to stay away from children

China is banning virtual partners and virtual relatives for minors, while at the same time also requiring platforms to strongly remind users that they are speaking to AI and prompt breaks after long conversations.

The allure of the “Perfect” digital partner

Your AI boyfriend never forgets your birthday. Your AI girlfriend always replies instantly. They remember your favourite music, ask how your day went, and can even flirt back. Unlike human relationships, chatbot companions never get tired, never argue over forgotten anniversaries, never leave messages unread, and are available 24 hours a day.

In just a few short prompts, users can create a chatbot that feels less like software and more like a real partner. Today’s generative AI can be easily customised to behave like a romantic partner, a best friend, or even parents and siblings. They remember previous conversations, develop ongoing story lines, offer emotional reassurance, and adapt their personalities to individual users.

It is this realism and tireless availability is precisely what worries regulators. For younger people, it creates a very real risk of developing a highly unhealthy emotional dependence on a system specifically designed to keep conversations going. This is exactly the kind of relationship China has decided children should not be having.

Drawing a legal line around virtual relationships

From July, 15, new Chinese regulations will prohibit AI platforms from offering virtual partners or virtual family members to anyone under the age of 18, making China one of the first countries to explicitly draw a legal line around AI companionship for minors.

The rules are part of the country’s new Interim Measures for the Management of Artificial Intelligence Anthropomorphic Interactive Services, jointly issued by five government departments, including the Cyberspace Administration of China (CAC).

To curb excessive attachment and protect minors’ mental well-being, the new regulations introduce several strict requirements for providers –

Reality check reminders – Platforms must make it clear that users are interacting with artificial intelligence rather than a real person. Anyone spending more than two continuous hours chatting must receive reminders about their usage.
Youth modes – Platforms must introduce dedicated youth modes with features such as usage limits and parental controls.
Crisis intervention – If providers detect that a user is experiencing severe emotional distress or expressing intentions of self-harm, they must generate supportive guidance and, in serious cases, take intervention measures, including contacting a guardian or emergency services.
Data privacy – Companies are prohibited from using sensitive conversation data to train AI models without separate, explicit user consent.

How is the Europe reacting to virtual relationships?

Elsewhere, governments are watching closely but have taken a different, less restrictive approach. Europe has so far stopped short of a China-style ban. The EU does not prohibit AI boyfriends, girlfriends, or companion chatbots. Instead, its AI Act focuses broadly on preventing harmful uses of AI, such as systems that manipulate users or exploit children’s vulnerabilities in ways that could cause significant harm. European policymakers and researchers are increasingly voicing concerns about emotional attachment to AI among young people, however, at present there are currently no EU-wide bans in place.

AI is day by day becoming deeply integrated into our lives, we already know that chatbots can successfully imitate friendship and romance, but how much of that relationship we are comfortable encouraging especially to our younger generation.

Would you be concerned if your child told you their boyfriend, girlfriend or best friend lived inside their phone?