MAS flags risks of using AI agents in digital-asset trading

MAS has cautioned about the risks associated with utilizing AI agents in digital asset trading, as these agents are capable of not only analysing data but also initiating and executing financial transactions. Chia Der Jiun, managing director of the Monetary Authority of Singapore, emphasised that as AI agents become more autonomous, they pose significant implications for liquidity and collateral management.

To ensure safe and effective deployment, he stressed the need for robust controls around authority, permissions, identity, accountability, and governance. MAS has already published the Safeguards for Agentic Finance at Runtime (SAFR) white paper and released an open-source reference implementation to facilitate industry experimentation.

The focus of regulatory discussions now shifts from simply verifying the effectiveness of technologies to ensuring operational safety, commercial viability, and scalability. The evolving regulatory landscape is increasingly collaborative, incorporating soft law instruments alongside traditional legislation, which has been noted as promising by panelists at the conference.