MNI BRIEF: Market Regulators Need To Look At AI Impact-Ex PBOC

Regulators likely need to address the outcomes and actions generated by black-box artificial intelligence models to regulate financial markets, even though the impact of AI on the formulation and implementation of monetary policy is not significant at present, said former governor of People’s Bank of China Zhou Xiaochuan said at Bund Summit on Thursday in Shanghai. 

Although the IT industry, along with emerging AI, helps in more accurately assessing situations and enables macro policy makers to respond more precisely, using machine learning to infer and predict financial instability based remains an important direction, he pointed out. 

There is an expectation that if financial institutions and activities employ AI, they should provide transparent and interpretable models to regulators, but the development of AI inevitably introduces black-box factors to models, making them difficult to interpret. What is more, if AI models extensively utilize short-term and high-frequency data, the outcomes are likely to conflict with the fundamental, long-term stability requirements essential for macro policy, he warned. (See MNI: Relaxed Policy To Drive PBOC Over Next Five Years - Advrs)