China launches first security standard for intelligent agents in financial sector

The Beijing Financial Technology Industry Alliance recently released the 'Security Requirements for Intelligent Agent Technology in Financial Applications' group standard, marking the first such standard in China focusing on the security of intelligent agents in financial apps. The standard mandates that third-party agents must obtain both user and institutional authorization to operate financial apps, a concept summarized as 'dual authorization'. Developed by the Beijing National Financial Technology Certification Center and major financial institutions and tech companies, the standard covers five key areas, including data security and risk control. It was released in response to incidents in December 2025, where users faced login and payment issues after a smartphone model with an AI assistant was launched. The standard aligns with national guidelines issued in May 2026, which called for stricter management of intelligent agent permissions and behaviors. Phone manufacturers like Huawei and Tencent have adjusted their technical approaches, requiring app developer consent for operations. New features such as A2A capabilities, introduced by WeChat and Honor in June 2026, enable second-level authorization beyond user consent, reflecting broader industry shifts toward enhanced security and compliance.