(Yicai) Oct. 24 -- Artificial intelligence technology is changing the global financial system, and due to the potential magnitude of these changes, it will likely reshape the operational models of the banking and financial sectors, according to the former governor of China's central bank, who also warned that AI is not suitable to make monetary policy decisions.

With the advancement of AI tech, the human-machine relationship in the financial industry is evolving from "human-led with machine assistance" to "machine-led with humans serving as customer interfaces," Zhou Xiaochuan said during a roundtable discussion themed "AI Governance and International Cooperation in the Financial Sector" at the seventh Bund Summit in Shanghai, which opened yesterday and ends tomorrow.

The massive amount of data accumulated by financial institutions has laid a solid foundation for deep learning, Zhou pointed out. Banks will become more reliant on AI in payment, pricing, risk management, marketing, and other areas, which may lead to significant changes in employee numbers, he said, adding that customer behavior will also change, with more and more of them getting used to interacting with machines.

He highlighted that AI tech will also transform financial regulatory models. For example, in anti-money laundering and counter-terrorist financing, the use of AI enables watchdogs to identify potential risks more efficiently, enhance their capabilities, and transform the traditional inefficient regulatory model that relies on manual detection, reporting, and handling of violations, he noted.

In addition, AI has enormous application potential in achieving financial stability, Zhou said, noting that the collapses of Silicon Valley Bank and Silvergate Bank happened very suddenly, and traditional regulatory indicators struggled to issue timely warnings. AI tech can be used to perform deep learning from historical financial data to predict, identify, and address unstable risk factors in the financial industry, he added.

Since financial system instability often erupts suddenly after a period of accumulation, AI models for predicting such risks need to process a large amount of structured and unstructured data, including multimodal information such as social sentiment and public opinion, Zhou pointed out.

However, AI is not capable of making monetary policy decisions because the process requires stable and continuous data observation and judgment, while AI's strength lies in high-frequency data processing capabilities, according to Zhou.

Editors: Tang Shihua, Martin Kadiev