UNICEF has officially congratulated the Chinese government for issuing new regulations designed to protect children from the risks associated with artificial intelligence. In a press release dated April 24, 2026, the United Nations agency praised the “Interim Measures for the Administration of Artificial Intelligence Anthropomorphic Interactive Services,” calling the policy a “pioneering” and “significant global step” toward regulating AI-powered emotional interaction services, such as virtual companionship and emotional care. The statement represents a notable endorsement from a major international body of China’s approach to AI governance in a domain that is attracting increasing attention from regulators worldwide.
The new measures introduce several unprecedented safeguards for the nearly 196 million children in China who are connected online. Most notably, the regulations establish a clear red line by explicitly banning virtual intimate relationships for children, including AI-based virtual family members or partners. UNICEF highlighted this specific provision as an “unprecedented safeguard in global AI policy discussions,” emphasizing the need to shield vulnerable populations from false intimacy and the potential psychological harms of emotionally immersive AI systems. The concern is not hypothetical: AI companion apps have seen explosive growth in China and globally, with some services specifically marketed to young users seeking emotional connection.
UNICEF China, in collaboration with the UNICEF Office of Research, provided technical input to help strengthen the child protection safeguards within the measures. This engagement reflects a broader effort to ensure that innovation in AI advances children’s rights and safety, and it gives the regulations a degree of international credibility that purely domestic rules might lack. The collaboration also demonstrates that China is willing to engage with international organizations on sensitive regulatory matters, at least in domains where the goals are broadly aligned.
(Related: China Bans AI Virtual Companions for Minors in New Digital Human Regulations)
Comprehensive Protections and Parental Controls
Beyond the ban on virtual intimacy, the measures explicitly prohibit AI services from generating content for children that could encourage unsafe behavior, extreme emotional responses, or harmful habits affecting physical or mental health. This provision addresses a growing concern about AI systems that use sophisticated emotional manipulation techniques to maximize user engagement, a practice that is particularly harmful when the users are minors.
The regulations also mandate age-appropriate protections, requiring developers to implement systems that enable parents and guardians to receive safety risk alerts, monitor usage patterns, block specific AI characters, and restrict in-service spending. This parental control framework recognizes that children’s digital safety cannot be guaranteed solely by platform-level restrictions; it requires active parental involvement, supported by accessible, easy-to-use tools. The requirement to provide these tools places a clear obligation on service providers to invest in safety infrastructure, not just content moderation.
Transparency is another key pillar of the new rules. The measures require that all AI-generated content be clearly labeled, ensuring that children understand they are interacting with an artificial system rather than a real person. This labeling requirement addresses a fundamental concern about the blurring of boundaries between human and AI interaction, particularly in emotionally charged contexts. Furthermore, child protection is embedded in the development process through mandatory security assessments that require the implementation of online protection measures for minors before a service can be deployed.
(Related: China Moves to Regulate Booming Digital Human Industry with New Draft Rules)
A Blueprint for Global AI Governance
UNICEF noted that these elements reflect a “comprehensive and innovative response that offers important reference points for child-centred AI governance globally.” While the agency acknowledged that further alignment with international child rights standards remains possible, it views China’s approach as a model for integrating child-centered safeguards into AI governance frameworks early. This is a significant statement: it positions China not as a laggard in AI governance, but as a potential leader in a specific and increasingly important domain.
The broader context is significant. As AI systems become increasingly sophisticated and emotionally engaging, the need for robust regulations to protect minors is becoming a pressing global issue. Many countries are still in the early stages of developing AI governance frameworks, and the question of how to protect children from the risks posed by emotionally immersive AI systems is largely unaddressed in most jurisdictions. China’s proactive stance, supported by international bodies like UNICEF, could influence how other nations approach this complex intersection of artificial intelligence and child welfare, potentially establishing a global standard for child protection in the age of AI.
The new measures also reflect a broader pattern in China’s approach to AI governance: moving quickly to regulate specific high-risk applications rather than waiting for comprehensive, cross-cutting AI legislation. China has already issued regulations covering algorithmic recommendations, deepfakes, and generative AI, and the new measures on anthropomorphic interactive services add another layer to this growing regulatory framework. This sector-by-sector approach allows regulators to address specific harms as they emerge, without the delays inherent in drafting broad legislation. While critics may argue that this piecemeal approach lacks coherence, it has allowed China to move faster than most Western countries in establishing concrete rules for AI-related risks. The UNICEF endorsement suggests that, at least in the domain of child protection, this approach is producing results that meet international standards.