周小川：货币政策是“慢变量”，过快响应可能引发不必要波动

Core Viewpoint - The financial industry is at a transformative crossroads due to the rise of artificial intelligence (AI), which is seen as a significant marginal change rather than just a technological tool [2] Group 1: Evolution of Banking - The nature of the banking industry has fundamentally changed over the past 60 to 70 years, evolving from traditional banking to a data processing industry [2] - Core banking functions such as payment, pricing, risk measurement, and marketing now heavily rely on data analysis and model computation [2] - The relationship between humans and machines has shifted from human-led to machine-led, with humans primarily acting as interfaces between machines and customers [2] Group 2: Impact of AI on Employment - The application of AI in banking is expected to significantly reduce the workforce size due to the reliance on data analysis and reasoning models [2] - Customer behavior has evolved, with more clients preferring machine interactions over human involvement, further driving AI's role in banking [2] Group 3: Opportunities and Challenges in Regulation - AI presents both opportunities and challenges in regulatory frameworks, particularly in areas like anti-money laundering, where machine learning can identify suspicious activities from vast datasets [3] - In financial stability, machine learning may help predict critical moments in the market, but it requires handling unstructured data and social sentiment [3] Group 4: Transparency and Policy Implications - The "black box" nature of AI models conflicts with the need for transparency in regulatory requirements, and over-reliance on short-term data can lead to misalignment with long-term financial stability [4] - AI's influence on central bank policies, particularly the dual-pillar framework of monetary policy and macroprudential regulation, requires further observation and research [4] - While AI can enhance data collection and analysis for monetary policy decisions, the slow nature of monetary policy adjustments necessitates a cautious approach to avoid unnecessary volatility [4]