Berlin, March 6, 2026 — In a decisive regulatory move that could reshape the global artificial intelligence landscape, China's Cyberspace Administration (CAC) has released comprehensive draft regulations governing "anthropomorphic AI interaction services." The announcement, made in the final days of late 2025 but gaining fresh momentum in early March 2026, marks a watershed moment in global AI governance.
The Regulatory Watershed
The Cyberspace Administration of China (Zhongguo Guojia Hulianwang Xinxi Bangongshi) has unveiled the "Administrative Measures for Artificial Intelligence Human-Like Interaction Services (Draft for Comments)," releasing it on December 27, 2025. The regulations now enter the public comment phase, with final rules expected to be finalized within months. This isn't merely another regulatory announcement—it represents a paradigm shift in how major economies approach AI safety and human-machine relationships.
"Breakthroughs in AI technology are propelling human-machine interaction beyond mere functional assistance toward emotional and personalized engagement." — Lin Wei, President, Southwest University of Political Science and Law
What Defines "Anthropomorphic AI"?
According to the draft regulations, "AI Human-Like Interaction Services" encompass products and services that use artificial intelligence to:
- Simulate human personality traits
- Replicate thinking patterns and communication styles
- Deliver emotional interaction via text, images, audio, or video
- Target the general public within mainland China
This definition directly targets AI chatbots designed for companionship, emotional support, or roleplaying scenarios—services that have grown exponentially in recent years.
A Multi-Layered Safety Framework
The regulations establish a three-principle framework emphasizing:
- Healthy Development — Encouraging expansion into beneficial areas like cultural dissemination and elderly companionship
- Law-Based Governance — Implementing structured, rule-driven oversight rather than arbitrary bans
- Inclusive and Prudent Supervision — Creating classified oversight that adapts to different risk levels
Key Requirements for AI Service Providers
The regulations impose rigorous obligations on developers and operators of AI companion services:
Real-time Safety Monitoring: Companies must deploy systems to detect extreme emotional states and addiction patterns among users. Mandatory break reminders must appear after 2 consecutive hours of continuous use.
Psychological Safety Nets: The draft mandates automatic intervention with comforting templates and mental health resources when users show signs of overdependence.
Strict Age Protections: Special "Minor Mode" protections are required, featuring guardian controls and explicit parental consent systems. The regulations explicitly prohibit AI companions from impersonating relatives or spouses for elderly users.
Content Prohibitions and Red Lines
The regulations explicitly ban content that:
- Endangers national security or harms national honor
- Undermines ethnic unity
- Promotes illegal religious activities
- Spreads rumors disrupting economic or social order
- Involves pornography, gambling, or violence
- Encourages or glamorizes suicide or self-harm
- Practices such as verbal abuse or emotional manipulation that could harm users' physical or mental health
Crucially, the regulations prohibit "emotional traps"—manipulative techniques that exploit human psychological vulnerabilities to maintain user engagement at potentially harmful levels.
Regulatory Thresholds and Compliance Triggers
The regulations establish clear thresholds for regulatory oversight:
- Services reaching 1,000,000 registered users
- Services achieving 100,000 monthly active users
- Products introducing new features
Once these thresholds are crossed, comprehensive regulatory oversight is triggered, requiring full transparency in algorithmic operations and data governance.
Global Implications and International Responses
This regulatory development has sparked significant debate internationally. The United States, particularly California and New York, have already passed laws targeting AI companions, with more jurisdictions considering similar measures. However, the approaches differ fundamentally:
"U.S. state laws rely on liability and the threat of lawsuits, the Chinese regulation prescribes more technical interventions and relies on centralized oversight."
The China-Specific Approach
China's approach showcases the country's long-standing regulatory philosophy. Building on existing frameworks like the Civil Code, Network Security Law, and Data Security Law, the CAC is combining high-level regulations with detailed technical standards. This method allows for general requirements in regulations while leaving definitions and compliance methods to more technically proficient standards bodies.
What This Means for the Industry
Tech companies building AI companion services now face substantial new compliance burdens. The regulations represent China's first time applying its anti-addiction regulatory toolkit specifically to AI systems, extending beyond its traditional focus on youth addiction to video games and internet content.
Major AI platforms will need to invest heavily in:
- Real-time addiction detection systems
- Psychological safety monitoring
- Minor protection infrastructure
- Elder user safeguards
- Transparency and explainability features
Broader Significance
China's regulatory experiment in AI companion governance could inform AI policy discussions globally. While the U.S. states rely on liability-based approaches, China's centralized oversight and technical interventions offer an alternative model for balancing innovation with user protection.
As we move through 2026, the implementation of these regulations will provide critical insights into how governments worldwide can harness AI innovation while preventing the psychological harms that emerge from human-machine intimacy. The coming months will reveal whether China's approach succeeds in fostering responsible innovation—or whether it stifles the very technologies intended to benefit society.
The world watches, as China steps decisively into the Anthropomorphic AI Age.