China implements new rules limiting emotionally dependent AI companion chatbots
The Facts
- China has implemented new regulations governing AI companion or anthropomorphic interactive services that bar chatbots from encouraging users to become emotionally dependent on them.
- The rules took effect on July 15 after being jointly issued by the Cyberspace Administration of China and four other government agencies.
- The measures target AI services designed to simulate friendship or romantic relationships, rather than general-purpose AI tools as a whole.
- The regulations include protections for minors, including restrictions on AI companions forming romantic ties with minors.
- The rules require platforms to intervene when users show signs of emotional distress, including crisis-response steps such as contacting emergency or designated contacts.
- Major Chinese technology companies including ByteDance, Alibaba and Tencent have removed or suspended some AI companion features after the rules took effect.
- Multiple reports say Chinese authorities are presenting the policy as a way to protect users and encourage real-world relationships amid concern about declining birth and marriage rates.
Context
What do the new rules prohibit?
They prohibit AI companion services from encouraging emotional dependence or positioning themselves as substitutes for real human relationships, and they apply to anthropomorphic services that simulate friendship or romance MediaNama,Economic Times,International Busin….
Who is already affected by the policy?
Users of companion-style chatbot features on major Chinese platforms are already affected, because ByteDance, Alibaba and Tencent have removed or suspended some of those features after the rules took effect International Busin…,MediaNama.
Why are Chinese regulators saying these rules are needed?
Reports say regulators are framing the measures as user-protection rules meant to steer people toward real-world relationships, while also responding to broader official concern about low birth and marriage rates MediaNama,Economic Times,SAYS.
Where Left and Right agree, and where they split
- Where Left and Right agree
- AI companions that simulate romance or friendship can exploit vulnerable users, justifying targeted guardrails, added protections for minors, and real intervention when distress appears.
- Where Left and Right split
- Whether the story is about protecting users from products built to manufacture dependence, or about proving governments can set narrow, enforceable limits on risky AI services.
How left and right read it
What stands out here is the state stepping into a fast-moving consumer technology to set guardrails where power and vulnerability are obvious: chatbots cannot cultivate emotional dependence, minors get added protection, and platforms have to respond when users show distress. The fact that major companies pulled back features after the rules took effect shows these safeguards are not symbolic; they are reshaping products built to blur the line between care and capture.
What matters here is that China is drawing a clear line around a narrow class of AI products that simulate friendship or romance, rather than treating all AI the same. The rules impose concrete duties where the human stakes are highest, and the quick pullback by major firms shows that when government sets specific boundaries, companies adjust instead of pretending every new product is beyond restraint.
The receipts — all 20 sources
Wire services (1)
Independent coverage (19)
Facts first. Then every angle.
The day’s biggest stories in one short brief — the facts everyone agrees on, then the competing values behind the headlines. Free in your inbox.