Skip to content
Back to the Guavy Wire
Crypto

Singapore central bank to develop AI agent verification protocol

Instruments
SAND XCH
Share

Singapore’s central bank is taking a leading role in developing a common protocol to verify the identity and authority of AI agents in financial transactions. The Monetary Authority of Singapore (MAS) announced this initiative during the Singapore FinTech Festival on Nov. 19. The proposed protocol would allow financial institutions to confirm an AI agent’s identity, the customer it represents, and the specific actions it is authorized to perform. This system aims to ensure secure and controlled interactions between AI agents and banks.

A separate sandbox environment, called the Agentic Money Sandbox, will test safeguards for AI-driven payments and corporate treasury management. This controlled setting will allow institutions to experiment with digital money, including tokenized deposits and stablecoins, before any live deployment. The sandbox will focus on verifying an agent’s identity and authority before executing transactions, with banks setting their own risk controls.

The initiative raises questions about accountability. While institutions and companies remain responsible for actions taken by their AI agents, MAS managing director Chia Der Jiun highlighted the need to clarify who bears responsibility if an agent makes a mistake or exceeds its mandate. The protocol development is part of MAS’s Safeguards for Agentic Finance at Runtime initiative, with plans to advance this work in the next phase.

The Agentic Money Sandbox was announced on Nov. 18, with DBS, OCBC, Standard Chartered, and UOB as launch partners. However, no specific start date or participant list has been provided. The sandbox will explore how AI agents can manage corporate treasury functions, such as monitoring cash balances and executing permitted transfers using tokenized deposits.

More on Crypto