Financial regulators address AI agent risks and banking modernization

Financial institutions are navigating the complexities of technological modernization and the integration of artificial intelligence. While many banks focus on speed and intelligence, industry experts note that modernization must prioritize dependability and control to avoid regulatory and operational consequences, particularly in sensitive areas like KYC, AML, and cross-border payments.

In response to the rise of AI agents in finance, the Monetary Authority of Singapore (MAS) has introduced a framework titled ‘Safeguards for Agentic Finance at Runtime’ (SAFR). Developed with major firms including HSBC, JPMorgan Chase, and Mastercard, the approach proposes checking an AI agent’s identity, permissions, and risk limits before it executes a transaction, rather than reacting to errors after the fact.

The SAFR architecture utilizes an identity layer, a rules repository, an action-checking engine, and an audit log to manage AI decisions. While MAS is moving toward concrete technical safeguards, other regulators, such as the United Kingdom’s Financial Conduct Authority, are currently adopting a more cautious stewardship-based approach to manage AI risks.