AI, Digital Cash Set To Reshape Financial Systems
KUALA LUMPUR, July 9 (Bernama) -- The emergence of artificial intelligence (AI)-driven autonomous agentic systems could become significantly more consequential for financial systems when paired with digital cash, such as stablecoins, tokenised deposits or deposit tokens.
In a research note today, Moody’s Ratings said that as agentic systems evolve from decision-support tools to autonomous economic actors, efficient participation across platforms will increasingly require interoperable, machine-readable payment infrastructure.
“Digital cash could enable agents not just to decide, but also to transact, settle obligations and coordinate value transfer more efficiently across digital environments,” it said.
The rating agency noted that digital cash has the potential to turn AI agents from tools that recommend actions into delegated economic actors that can initiate payments, move liquidity, post collateral and settle obligations.
It said the authority of AI agents remains constrained by mandates, limits, compliance checks and settlement rules.
Moody’s added that AI agents with digital cash have the potential to unlock several use cases for institutions, including automating treasury management, collateral posting, capital markets settlement, trade finance, supply chain and cross-border payments, as well as e-commerce workflows.
“Stablecoins, tokenised deposits and deposit tokens give agents continuous access to value and interoperability, while deterministic rules govern authorisation, compliance and settlement.
“As more assets become tokenised, agents will be able to coordinate entire financial processes in real time,” it said.
Moody’s Ratings also highlighted that AI agents expand the risk surface across financial, technological, operational and regulatory dimensions.
It said that giving agents access to digital cash increases financial, cyber, operational, regulatory, and legal risks, including liquidity runs, correlated agent behaviour, prompt injection, application programming interface (API) compromise, authorisation gaps, and unclear liability.
“As such, strong controls, including know-your-agent frameworks, transaction limits, audit trails, real-time monitoring and human override mechanisms, will be essential,” it added.
-- BERNAMA