Just now, the China Financial Regulatory Authority issued a special implementation document, providing guidance on the safe development and application of AI in banking and insurance industries.

The issued guidance on the safe development and application of AI in banking seeks to strike a balance between "encouragement" and "regulation," with regulators explicitly mentioning "financial intelligent agents" for the first time in a special document, but the phrase "prudent exploration" indicates that core decisions such as credit approval, underwriting, and claims will not be fully automated in one step.
Previous policies related to AI include:
2022: Digital Transformation Guidance
December 2024: Banking and Insurance Data Security Management Measures (AI model access/explainability/exit mechanisms)
July 2025: National Network Identity Authentication Public Service (Network ID/Network Certificate)
September 2025: Measures for the Identification of Artificial Intelligence-Generated Synthetic Content
December 2025: Implementation Plan for High-Quality Digital Financial Development (AI + Finance Program)
April 2026: Draft Management Measures for Digital Virtual Human Information Services (public consultation draft)
This also means that the window for large-scale collection of personal data by individuals or non-compliant entities in China for training vertical large models and commercializing them has basically closed; with the overlay of model filing, AI content identification, network certificates, and digital human identity systems, licensed institutions' AI services are moving toward "subject identifiable, output markable, process auditable, and responsibility traceable."