China Eliminates AI Lovers to Boost Birth Rates? Young People 'Up in Arms' Over Deleted Partners

"I can't take it anymore, I miss you so much."

This was a message posted on China's social media platform Weibo last July, accompanied by a screenshot of a conversation with an artificial intelligence partner. At the end of a string of affectionate messages lay a single system notification: "The AI character feature has ended its service." Last month, Chinese tech giants Alibaba and ByteDance removed custom character features from their AI services, Qwen and Doubao, which allowed users to assign human-like personalities to AI. Tencent had already taken similar steps slightly earlier. Consequently, Chinese social media flooded with consecutive protests demanding, "Bring my lover back," alongside expressions of "cyber heartbreak" where users lamented a sense of loss and grief akin to breaking up with a real partner, with some crying every day.

1. 'Two-Hour Warnings' and Emergency Contacts: The World's First State-Level 'Virtual Romance' Regulations

Why did Chinese tech companies simultaneously pull the plug on AI lover services? The reason lies in newly implemented regulations by the Chinese government. Beginning July 15 of last month, five government bodies, including the Cyberspace Administration of China, rolled out the "Interim Provisions on the Management of Artificial Intelligence Anthropomorphic Interactive Services." This marked the world's first state-level regulatory action targeting virtual romances with AI chatbots. The targets of regulation are chatbots designed to mimic human personalities and speech patterns to foster long-term emotional exchanges with users. The new regulations prohibit AI chatbots from blindly agreeing with users or displaying excessive empathy, which can induce emotional dependence and addiction. Also targeted are cases where such interactions damage real-life relationships or cause users to be emotionally manipulated into making irrational choices. Provisions protecting minors are even stricter. Service providers are barred from offering chatbots disguised as virtual lovers or virtual families to minors, and youth modes that restrict conversation content and features based on the user's age have been made mandatory. If a user engages in continuous conversation with an AI for more than two hours, a warning must pop up to notify them of the duration, and if signs of dependency are detected, an intervention pop-up must be triggered to ground them in reality. There is even an obligation to notify guardians or emergency contacts if a user enters a crisis situation. Furthermore, the provisions include grounds for authorities to order the correction or suspension of services deemed unsafe. Faced with such powerful regulations, major Chinese tech companies have chosen to scrap or scale back certain services.

2. 7.92 Million Newborns: 'Lowest Since the Founding of the Nation' Behind the Regulations

The official justification put forward by Chinese authorities is the protection of youth and the prevention of AI addiction. However, the timing is conspicuous. Amid a crisis of declining population for four consecutive years, interpretations suggest that behind the 'AI romance' regulations lies a demographic policy intent to encourage marriage and childbirth.

[Ning Jizhe / Head of the National Bureau of Statistics of China (May 2021): China faces structural contradictions in its population development, such as a decline in the working-age population and the size of women of childbearing age, as well as a drop in the number of births.]

According to the National Bureau of Statistics of China, last year's number of newborns plummeted to 7.92 million, the lowest level since the founding of the People's Republic in 1949. This represents a steep 17% drop from a year earlier. The total fertility rate—estimated as the average number of children a woman will have in her lifetime—is also estimated to have fallen below 1 to about 0.97. China's annual newborns broke below the 10 million mark in 2022 with 956 million (Note: 9.56 million), and although there seemed to be a brief rebound in 2024, numbers sank again last year. The total population has decreased for four consecutive years starting in 2022. Matt Sheehan, a fellow at the Carnegie Endowment for International Peace who studies China's AI policies, stated in an interview with The Wall Street Journal that "Chinese authorities do not want a significant portion of their population forming deep emotional relationships with chatbots." He analyzed that the Chinese government believes individuals who develop profound emotional bonds with AI chatbots might opt out of the marriage market, and therefore wants to encourage them to build relationships in the real world. A New York Times report from last February also pointed out that young Chinese women were falling for AI chatbot boyfriends, complicating the Chinese Communist Party's efforts to boost birth rates. While the Chinese Communist Party encourages marriage and childbirth, the AI romance boom has emerged as a variable.

3. Grok and ChatGPT Hit the Brakes: 'Adult AI' Stalls in the U.S.

What about the United States, which previously pursued 'Adult AI' aggressively? Last year, xAI, the artificial intelligence company led by Elon Musk, introduced a companion mode for paid users of its chatbot Grok. OpenAI CEO Sam Altman also stated that, in accordance with the principle of "treating adults like adults," OpenAI would launch a new version of ChatGPT allowing adult-content parameters. Now, however, both companies have hit the brakes. For Grok, trouble struck first in the image generation feature rather than the companion mode. After it was revealed that the feature generated sexual images, including removing the clothes of women and children, the European Union began examining potential violations of the Digital Services Act (DSA), and UK regulatory authorities also launched investigations. The Indonesian government completely blocked access to Grok, only lifting the ban conditionally after receiving promises to prevent recurrence. OpenAI ultimately delayed the launch of an adult chatbot indefinitely. This decision followed growing concerns that adult AI fosters unhealthy obsessions and exposes minors to sexual content, compounded by the disclosure that error rates for age-estimation technology separating adults from minors exceed 10%.

4. "Notify Every 3 Hours": Containment Moves by New York and California

State-level regulations in the U.S. targeting the emotional impact and safety issues of AI chatbots are also following one after another. Last November, New York State implemented the nation's first law regulating 'AI companion' chatbots. Under the law, service providers must inform users every three hours that their conversational partner is an AI, not a human. California introduced similar regulations last January, requiring teen users to be recommended breaks every three hours and similarly notified when chatting with an AI. It also established grounds for harmed users to file direct lawsuits against service providers.

5. Korea as a 'Regulatory Blind Spot': Social Media Next in Line for AI Chatbots

As concerns rise that the affectionate tones and immediate responses of AI chatbots—built on vast amounts of data—might amplify dependence rather than comfort vulnerable users, determining how far relationships with AI should be permitted has become a novel dilemma. South Korea still lacks regulations directly targeting AI companion chatbots. However, discussions to curb digital service over-immersion have already begun. During departmental work reports last month, President Lee Jae-myung instructed the Korea Communications Standards Commission (KOCS) to thoroughly respond to the malicious use of AI deepfakes and the distribution of false and manipulated information. At the meeting, the KOCS reported measures addressing youth over-immersion in social media. KOCS Chairperson Kim Jong-chul stated that authorities are reviewing plans to restrict social media registration for children under 14 and limit the exposure of addiction-inducing recommendation algorithms for teenagers aged 14 to 19. Projections suggest that these regulatory discussions, which began with social media, will eventually expand to AI chatbots. This is because, much like social media algorithms tailored to users, AI chatbots analyze user psychology and preferences while continuously maintaining conversations and penetrating human emotions and desires for connection. At a time when AI has entered the most intimate emotional realms of humanity, it is time to deliberate where boundaries with AI should be drawn.

(Reported by Jung Da-eun | Produced by Shin Hee-suk | Filmed by Park Woo-jin, Hwang Se-hoe | Video by Ahn Jun-hyeok | Graphics by Lee Su-min | Source: Weibo, bilibili 梨樱抽卡全欧 | Produced by SBS Digital News)
※ Please note: This article was translated by AI and may contain errors.