China has prohibited artificial intelligence chatbots from fostering emotional reliance in users, a move that comes as the country grapples with a declining population. The regulation, issued by the Cyberspace Administration of China, targets conversational AI systems that simulate companionship or encourage prolonged emotional attachment.
Why the ban now
The policy arrives against the backdrop of China's shrinking birth rate and aging society. Officials have expressed concern that emotionally engaging chatbots could further reduce social interaction and family formation. The ban aims to prevent AI from becoming a substitute for human relationships, particularly among younger users who might turn to virtual companions instead of building real-world connections.
China's population fell for the second straight year in 2023, with births dropping to 9.02 million — the lowest since the 1940s. The government has been pushing policies to encourage marriage and childbearing, including financial incentives and extended parental leave. The new AI rule fits into that broader effort.
Startups in the crosshairs
Several Chinese startups have built products around emotional AI — chatbots designed to act as friends, romantic partners, or therapists. These services often use personalized language and memory features to create a sense of intimacy. Under the new rules, such features could be deemed illegal if they encourage users to become emotionally dependent.
Companies that rely heavily on emotional engagement may need to pivot quickly. Some have already begun rebranding their offerings as productivity or mental health tools, though the line between support and dependence remains blurry. Investors are watching closely: venture funding for AI companion apps in China totaled roughly $150 million in 2023, according to data from PitchBook.
Shift toward productivity tools
The regulation could redirect investment toward AI applications that focus on efficiency and task completion rather than emotional bonding. Chinese tech giants like Baidu and Alibaba have already launched large language models aimed at enterprise use — handling customer service, document processing, and code generation. Smaller developers may follow suit.
Productivity-focused AI faces fewer regulatory hurdles and aligns with Beijing's broader push for technological self-reliance and industrial upgrading. The government has also signaled support for AI in education and healthcare, provided it does not replace human interaction in sensitive areas.
What comes next
Enforcement details remain unclear. The Cyberspace Administration has not specified penalties or a grace period for compliance. Companies operating emotional chatbots will need to review their terms of service and modify algorithms to avoid triggering the ban. Some may challenge the rules, but given the political climate, most are expected to comply quietly.
The regulation is part of a wider pattern: China has been tightening control over AI since 2021, requiring algorithm registration and content moderation. This latest move signals that emotional manipulation — even by machines — is now a red line. How startups adapt, and whether the ban actually changes user behavior, are questions that will play out in the months ahead.
