According to the White Paper on the Development and Security of Financial Artificial Intelligence (2025), The global financial AI market exhibits diverse growth trajectories: the United States leads technological innovation through market-driven approaches, the European Union establishes a robust risk-tiered regulatory framework via the Artificial Intelligence Act, while China expands its presence through policy guidance and vast application scenarios. Industry AI investment reached RMB 19.694 billion in 2024, projected to double to RMB 41.548 billion by 2027. Technologically, core applications are evolving from general-purpose large models towards financial-specific models, with cost reductions and localisation demands driving adoption. Applications have permeated front-, middle-, and back-office functions, forming an end-to-end intelligent system (front office: smart marketing, intelligent customer service, etc.; middle office: intelligent risk control, credit approval, etc.; back office: process automation, smart auditing, etc.).

Security Risks

Financial AI security risks are multidimensional: at the technical level, vulnerabilities such as adversarial attacks and backdoor attacks persist; at the data level, challenges include training data contamination, privacy breaches, and cross-border compliance; at the operational level, systemic risks arise from algorithmic bias and ‘model convergence’; at the governance level, challenges include blurred accountability boundaries and regulatory lag.

Governance requires dual-track ‘governance + technology’ drivers: Top-level design establishes trustworthy AI principles like fairness, transparency, and accountability, establishing cross-departmental governance committees; technical safeguards build full lifecycle security systems, using zero-trust architecture and privacy-enhancing technologies to strengthen defences, while AI-empowered security operations boost threat response efficiency.

Governance and Policy Recommendations

A diverse regulatory landscape has emerged globally, with China evolving from macro planning towards specialised regulations, moving towards systematic and agile oversight. Financial institutions are advised to incorporate AI security into their top-level strategies, establishing specialised governance frameworks and technical protection systems. Regulatory bodies should expedite unified management measures, promote regulatory sandboxes, and foster industry collaboration to build a secure ecosystem, ensuring the safe, trustworthy, and sustainable development of financial AI.