China AI Law Brief

Analysis of China’s AI Regulatory Developments 

Issue No. 1 | July 2026 

Prepared by Sarah Zhao
Partner, Rimon Law 

I. Executive Summary

Three major AI regulatory developments taking effect in July 2026 signal a significant shift in China’s AI governance approach. They establish a comprehensive regulatory framework consisting of one voluntary guideline (TC260-005) and two binding regulatory instruments covering AI ethics, AI agents, and anthropomorphic AI interaction services, including AI-powered chatbots and virtual companions. These new developments demonstrate China’s latest effort to translate broad AI principles into concrete, actionable rules, adding further regulatory clarity as the industry enters a new phase of compliance obligations. 

II. Background

The three AI regulatory developments discussed in this issue did not emerge in isolation. They reflect the convergence of rapid technological advances, evolving societal risks, and China’s broader strategy to strengthen AI governance. These developments help explain why China is moving from principle-based AI governance toward a more operational regulatory framework. 

Emerging Security Risks Posed by AI Agents 

During late 2025 and early 2026, open-source AI agent technologies, most notably OpenClaw, experienced rapid global adoption. Unlike traditional large language models that primarily generate responses, AI agents can autonomously perceive, plan, and execute tasks. Publicly reported OpenClaw-related incidents and security analyses highlighted emerging risks, including credential theft through malicious plug-ins, enterprise data leakage resulting from autonomous access to shared files, and indirect prompt injection attacks that manipulated agents into executing unauthorized internal commands. These incidents exposed governance challenges extending well beyond traditional content moderation and conventional cybersecurity controls. 

2, New Challenges Posed by Anthropomorphic AI 

At the same time, AI companion products, including virtual partners, emotional chatbots, and digital avatars, experienced rapid user growth. Their increasingly human-like interactions raised concerns over emotional dependence, addictive behavior, and psychological harm, particularly among minors and elderly users. 

These developments also exposed important regulatory gaps. Existing Chinese AI regulations, including the Interim Measures for Generative AI Services, primarily focus on content safety, algorithm governance, and data protection. They provide only limited guidance on emotional interaction, AI ethics, and the governance of AI agents. 

China’s Regulatory Response 

Against this backdrop, China accelerated implementation of its broader national AI strategy. The State Council’s 2025 Opinions on Deeply Implementing the “AI+” Action have emphasized that AI development should be safe, dependable, and controllable. Meanwhile, AI governance continues to evolve internationally, with the phased implementation of the EU AI Act and new legislation in California, New York, and other U.S. jurisdictions addressing AI companion technologies. 

These three AI regulatory developments represent complementary components of China’s regulatory response. TC260’s ethics and technical standards provide practical compliance guidance; the Interim Measures for the Administration of Anthropomorphic AI Interaction Services address the unique risks associated with highly human-like AI systems; and the Implementation Opinions on the Standardized Application and Innovative Development of Intelligent Agents establish a policy framework for the governance of autonomous AI systems. 

III. Analysis of the New Rules 

A. Rule on AI Ethics and Standards

The first AI regulatory development concerns TC260’s continued efforts to operationalize AI ethics and personal information protection through technical standards and implementation guidance. 

TC260’s Ethics-Safety Guidelines for Artificial Intelligence Applications 1.0, effective July 1, 2026, set out nine core ethics principles, including “enhancing human welfare,” “respect for life,” and “ensuring controllability and trustworthiness.” The Guidelines emphasize a human-centric approach, requiring AI to promote human well-being while preserving human autonomy. They also provide practical guidance on the respective responsibilities of developers, service providers, and end users. 

The TC260 framework also includes two national standards (GB/T): 

• Data Security Technology – Requirements for the Transfer of Personal Information Based on Personal Requests 
• Data Security Technology – Requirements for Personal Information Protection Compliance Audits 

Although these standards are not legally binding, they establish an important technical baseline for personal information portability and compliance audit methodologies. Companies and organizations processing personal information in China should review their internal procedures against the new standards, while audit institutions should ensure that their methodologies and capabilities align with the applicable requirements. 

Why It Matters 

Although not legally binding, these standards reflect regulatory expectations and are likely to influence future compliance audits, regulatory inspections, and enforcement practices. More importantly, they move China’s AI and personal information governance from broad policy principles toward more practical and operational compliance standards. 

B. Rule on Anthropomorphic AI

The second AI regulatory development concerns China’s first dedicated regulatory framework governing AI-driven emotional interaction, virtual companions, and chatbots simulating human personality. The Interim Measures for the Administration of Anthropomorphic AI Interaction Services, effective July 15, 2026, establish this framework. 

Key Obligations 

Service providers are required to: 

1. Complete algorithm filing with the CAC. 
2. Complete the required security assessment. 
3. Clearly disclose that the service is AI-generated rather than human. 
4. Implement addiction-prevention measures. 
5. Refrain from providing minors with virtual intimate relationship services, including virtual relatives or virtual life partners. Providing other anthropomorphic AI interaction services for minors under the age of fourteen requires the consent of a parent or legal guardian. 

The Measures apply to AI chatbot and companion providers, virtual influencer platforms, digital assistant developers, and gaming companies offering AI-driven NPC interactions. Products providing simulated emotional or personality-based interaction are likely to fall within their scope. 

Key Provisions 

• Prohibition on providing minors with virtual intimate relationship services, including virtual relatives and virtual life partners. 
• Enhanced protection for elderly users through mandatory risk disclosures and timely responses to requests for assistance. 
• Mandatory safety assessments once a service reaches one million registered users or 100,000 monthly active users. 
• Crisis intervention measures where users exhibit signs of self-harm or extreme psychological distress. 
• User interaction data may not be used for model training without explicit consent, and users must be able to delete their chat history. 
• Mandatory reminders after two hours of continuous use, together with an unobstructed exit mechanism. 

Why It Matters 

The Measures represent China’s first dedicated regulation addressing emotional AI companions. Although introduced after state-level legislation in the United States, including California’s SB 243 and New York’s Article 47, China’s framework adopts a different regulatory approach by: 

• protecting both minors and elderly users, whereas current U.S. state laws primarily focus on children; 
• prohibiting providers from offering virtual intimate relationship services, including virtual relatives and virtual life partners, to minors rather than relying primarily on age verification; 
• requiring crisis intervention rather than optional hotline referrals; and 
• establishing clear compliance thresholds based on user numbers. 

U.S. states such as California, New York, and Washington pioneered the first binding rules in this area, while China is establishing a national regulatory framework. The United States has not yet adopted comparable federal legislation. The differences primarily reflect distinct governance structures rather than one regulatory system being ahead of the other. 

Companies and organizations operating in China should assess whether existing or planned products fall within the scope of the Measures and, if so, review filing, security assessment, and related compliance requirements promptly. 

C. Rule on AI Agents

The third AI regulatory development introduces China’s first dedicated policy framework for AI agents as a regulatory category separate from generative AI. The CAC, NDRC, and MIIT jointly issued the Implementation Opinions on the Standardized Application and Innovative Development of Intelligent Agents, effective July 15, 2026, which establish this framework. 

The Opinions define AI agents as intelligent systems capable of autonomous perception, memory, decision-making, interaction, and execution. AI agents used in sensitive sectors, including healthcare, transportation, media, and public safety, face mandatory filing, compliance testing, and product recall provisions. The framework also targets 70% adoption of intelligent agents in smart terminals by 2027. 

Key Features 

• Defines AI agents as systems capable of autonomous perception, memory, decision-making, interaction, and execution. 
• Balances security, standardization, and innovation. 
• Identifies nineteen application scenarios spanning scientific research, manufacturing, healthcare, and consumer services. 
• Establishes a differentiated governance framework under which higher-risk sectors are subject to stronger regulatory oversight. 

Why It Matters 

The Opinions signal that China is beginning to regulate AI agents as a distinct governance category rather than simply treating them as an extension of generative AI. More importantly, they suggest that future regulation will increasingly focus on the deployment and governance of autonomous AI systems in higher-risk sectors. 

IV. Key Compliance Takeaways

Companies and organizations operating in China should focus on the following near-term priorities: 

1. Review personal information transfer and compliance audit procedures against the new TC260 standards. 
2. Assess whether existing or planned products may fall within the scope of the Anthropomorphic AI Interaction Services Measures. 
3. Evaluate whether existing AI products could be classified as AI agents under China’s emerging regulatory framework. 
4. Monitor additional TC260 standards, implementing guidance, and enforcement developments. 

These developments indicate that China’s AI governance framework is entering a more operational phase, with increasing emphasis on implementation, accountability, and the real-world deployment of AI systems. 

About the Author: Sarah Zhao is a Partner at Rimon Law, where she advises multinational companies and other global businesses on AI, data protection, privacy, cybersecurity, cross-border data transfers, technology transactions, and regulatory compliance. Fluent in both English and Chinese, she combines a deep understanding of different legal systems and cultural subtleties with experience advising clients on complex cross-border legal and compliance issues. 

China AI Law Brief is a periodic publication providing practical analysis of significant developments in China’s AI regulatory framework and their implications for companies, industry associations, government agencies, and international organizations. 

(This publication is provided for informational purposes only and does not constitute legal advice. Readers should consult legal counsel regarding their specific circumstances.) 

© 2026 Sarah Zhao, Partner, Rimon Law. All rights reserved.