AI-generated image
Artificial intelligence is poised to reshape the global financial system, changing the operational models of banks and other financial institutions, but it is not suitable for making monetary policy decisions, according to Zhou Xiaochuan, former governor of the People’s Bank of China.
Speaking at a roundtable discussion on “AI Governance and International Cooperation in the Financial Sector” at the seventh Bund Summit in Shanghai, Zhou highlighted the transformative role AI is playing across the financial industry. The summit, which opened yesterday and runs through tomorrow, brought together regulators, industry executives, and academics to explore how AI can be applied in finance while maintaining stability and regulatory oversight.
Zhou said that AI is altering the human-machine relationship within financial institutions. The traditional model, “human-led with machine assistance,” is gradually shifting to “machine-led with humans serving as customer interfaces.” Financial institutions are increasingly relying on AI for payment processing, pricing, risk management, and marketing, which could reshape staffing structures and change how customers interact with services.
“The massive amount of data accumulated by banks and financial firms provides a solid foundation for deep learning,” Zhou noted. He emphasized that AI can transform financial regulatory models, improving efficiency and risk detection. For instance, in anti-money laundering (AML) and counter-terrorist financing, AI enables regulators to identify potential risks faster and more accurately than traditional manual processes, which often rely on inefficient reporting and intervention.
Zhou also highlighted AI’s potential for supporting financial stability. The sudden collapses of Silicon Valley Bank and Silvergate Bank revealed limitations in traditional regulatory indicators, which failed to provide timely warnings. AI, through deep learning and predictive modeling based on historical financial data, can help regulators and institutions detect unstable risk factors before crises escalate.
“Financial instability often erupts suddenly after a period of accumulation,” Zhou said, noting that AI models need to handle vast amounts of structured and unstructured data, including multimodal information such as social sentiment and public opinion, to effectively forecast potential threats to system stability.
Despite these advantages, Zhou cautioned that AI is not suited for monetary policy decisions. Implementing monetary policy requires stable, continuous observation and judgment over economic conditions, while AI excels at processing high-frequency, large-scale data rather than making nuanced macroeconomic decisions.
Zhou’s remarks underscore the dual nature of AI in finance: while it offers significant efficiency gains, predictive power, and operational transformation, it cannot replace human judgment in areas requiring discretion, stability, and strategic foresight. Regulators and financial institutions are therefore tasked with balancing AI adoption with careful oversight to ensure that the technology enhances rather than destabilizes the financial ecosystem.
As AI continues to evolve, Zhou suggested that international cooperation and governance frameworks will be critical to harness its benefits while mitigating systemic risks, particularly given the increasingly interconnected nature of global financial markets.
