周小川：AI对货币政策影响尚不明显，金融稳定领域应用潜力更大

Core Insights - The former governor of the People's Bank of China, Zhou Xiaochuan, emphasized the need for cautious and sustained judgment in monetary policy, indicating that AI's high-frequency data processing capabilities do not align well with this requirement [1][3] - Zhou noted that while AI and machine learning excel in data collection and pattern recognition, their impact on monetary policy remains limited due to the slow-variable nature of monetary policy adjustments [3] - He highlighted the significant potential for AI applications in financial stability, particularly in predicting risks associated with sudden financial institution failures, which traditional indicators may not adequately forecast [3][4] Group 1 - Zhou Xiaochuan stated that monetary policy is inherently a "slow variable," adjusting in response to economic cycles or macroeconomic indicators, which do not change rapidly [3] - He pointed out that AI's ability to process high-frequency data does not match the need for stable and long-term judgment in monetary policy [3] - The potential for AI to analyze historical financial data and changes in the health of financial institutions to predict instability risks is considered a crucial direction for development [3] Group 2 - Zhou raised concerns about the "black box model" issue associated with AI, where the use of complex deep learning models by financial institutions could lead to challenges in regulatory oversight and risk management [4] - He mentioned that the high-frequency short-term data analysis provided by AI may not align with the long-term stability and fundamental orientation required by central banks [4]