The risk of “herding” in trading platforms due to the use of AI models is still low as financial institutions (FIs) in Singapore are still at a nascent stage of using AI models for decision-making, says the Monetary Authority of Singapore (MAS).
Herding is where individuals make similar decisions due to similar trained AI models or similar data aggregates.
The issue was brought up in a parliamentary question on the potential for conflict of interest when a trading AI platform system is considering both the platform’s interest and the customers’ interests, especially if without disclosure to customers.
Lawrence Wong, Deputy Prime Minister and Minister for Finance, and Chairman of MAS responded saying that it requires regulated FIs to have controls in place to avoid or mitigate conflicts between their interests and those of their customers. This approach is technology-neutral and is applied across all regulated FIs.
He added,
“Nevertheless, MAS continues to carefully monitor and assess the risks brought about by AI in financial markets. In recent years, MAS has worked with the industry to develop a framework which enables FIs to evaluate their AI systems against the FEAT Principles. MAS will continue to engage the industry on the responsible use of AI in finance.”
The regulator had introduced the FEAT Principles as early as 2018 to guide FIs’ responsible use of AI and data analytics in their products and services. The acronym stands for Fairness, Ethics, Accountability, and Transparency.