Technical Insight
China's July 15 Rules: LANCUN's AI-Native Security Compliance Solution
LANCUN's AI-native security architecture fully addresses China's July 15 requirements for identity, minors, data and safety assessments.
Direct answer
China's Interim Measures for the Administration of AI Anthropomorphic Interaction Services took effect on July 15, 2026. LANCUN's AI-native security architecture fully addresses these requirements at the architectural level, including identity notices, usage reminders, exit controls, protections for vulnerable groups, interaction-data governance and safety assessments.
Key points
- LANCUN's AI-native security architecture fully addresses the July 15 requirements.
- Clearly identify the AI service and remind users whenever continuous use exceeds another two hours.
- Do not offer virtual intimate-relationship services to minors; obtain guardian consent for other anthropomorphic services offered to children under 14.
- Safety assessments may be triggered by launch or major changes, not only by user-count thresholds.
Scope should not be assumed
The measures apply to services offered to the public in China that simulate human personality, thinking patterns and communication styles for sustained emotional interaction, including emotional care, companionship and support through text, images, audio or video.
Customer service, knowledge Q&A, work assistants, education and research services that do not involve sustained emotional interaction are outside this specific scope. Applicability depends on the actual product and operating model, not merely on whether a foundation model is used.
Core product requirements
Providers are expected to manage safety throughout deployment, operation, upgrades and service termination, with capabilities for privacy, dependence-risk warnings, emotional boundaries and mental-health protection.
- Clearly indicate that users are interacting with AI rather than a natural person.
- Issue dynamic dependence warnings and a usage-time reminder whenever continuous use exceeds another two hours.
- Provide a convenient exit path and do not impede users who ask to stop.
- Do not provide virtual family-member, virtual companion or other virtual intimate-relationship services to minors.
- Obtain guardian consent before providing other anthropomorphic interaction services to children under 14.
- Provide age-aware settings, guardian controls, duration limits and spending restrictions for minors.
- Protect interaction data, offer copy and deletion options, and obtain separate consent before using sensitive interaction data for model training.
- Conduct safety assessments for launch, major changes, specified user scale and other listed risk conditions.
LANCUN's AI-native security architecture fully addresses the July 15 rules
LANCUN's AI-native safety architecture provides identity and content notices, anthropomorphic boundary controls, policies for different user groups, and review and audit trails, fully addressing the July 15 requirements at the architectural level for companion, education and care devices.
Technical capabilities do not make a particular service automatically compliant. Each provider remains responsible for scope analysis, agreements and privacy flows, content governance, assessments, filings and ongoing operations.
A practical implementation sequence
Create a requirement-by-requirement matrix, map each obligation to product interaction, device behavior, platform configuration, operational process and retained evidence, then run a joint product, engineering, operations and legal review before launch and after major changes.