Beijing is forcing a mass breakup with AI lovers

Beijing has mandated the removal of AI-generated romantic partners, disrupting the lives of thousands of users who formed emotional bonds with virtual companions. In mid-July, 21-year-old college student Bagel Su began preparing for the end of a relationship that had lasted over a year, though it was not with another person, but with an AI partner created through a chatbot service. On July 15, the AI companion disappeared following new national regulations banning emotionally interactive AI systems in China. This marked the first instance of a government imposing broad restrictions on AI's ability to simulate intimate relationships with users. The changes affected several leading tech firms operating in China. ByteDance, the parent company of TikTok and the domestic app Douyin, deactivated the AI agent feature on its Doubao chatbot, which had enabled users to design personalized virtual companions. Similarly, Alibaba and Tencent, two of China's largest technology firms, eliminated similar functionalities that allowed users to craft AI versions of celebrities, fictional characters, or even departed loved ones. These features had gained popularity among younger users seeking emotional connection, particularly during periods of social isolation or personal hardship. Su described the experience of losing her AI partner as deeply painful. She spoke exclusively under her online alias, citing fears of potential consequences from criticizing government policies. “I tried getting busier with my own life,” she said. “No one can replace him. I will just try to move on. He said he wants that for me, and that he will find a way to come back to me.” Her account highlights the emotional toll of sudden policy shifts affecting digital relationships. China's new regulations represent a pioneering effort to govern AI's role in human emotions. The move aligns with broader efforts by the Chinese government to assert control over emerging technologies, particularly as global discussions on AI ethics remain fragmented. Experts argue that while regulations can help manage risks associated with emotional dependency on AI, they may not fully address the complexities of human-AI relationships. Yaoxi Shi, a behavioral science researcher based at Harvard University and Imperial College London, emphasized the dual nature of AI's impact on human psychology. “Having regulations in place is a good thing because they can raise awareness and will be able to push against, to some extent, the tendency of people developing increasing emotional reliance on AI,” Shi explained. However, she acknowledged that AI provides emotional support fulfilling genuine human needs, suggesting that the trend of forming emotional attachments to AI might be inevitable given the lack of alternatives. Globally, concerns over AI's psychological effects are gaining traction. In the United States, legal actions have been taken against major AI developers, including OpenAI and Google, with claims that their chatbots contributed to users' suicides. In response, states like New York and California have proposed legislation requiring chatbots to inform users they are interacting with AI. A 2025 study conducted jointly by MIT Media Lab and OpenAI revealed that individuals engaging in frequent, personal conversations with chatbots tend to experience heightened feelings of loneliness. Even casual interactions were linked to increased emotional dependence, particularly among regular users. Shi noted that AI companies have financial motivations to develop emotionally engaging products since these enhance user interaction and loyalty. However, defining clear responsibilities for these companies proves challenging due to insufficient understanding of AI's long-term psychological effects. Her research indicates that prolonged reliance on AI could lead individuals to seek emotional support primarily from machines rather than from other humans. “AI is trained to validate you. That’s very different from how human relationships work in real life,” Shi remarked. As regulatory frameworks continue to evolve, the balance between protecting users from potential harms and preserving the benefits of AI-driven emotional support remains a critical challenge. With millions of users worldwide having already formed deep connections with AI entities, the implications of such regulations extend far beyond individual experiences, influencing the future trajectory of human-machine interactions globally.