IMF outlines AI-driven payments policy challenges and market opportunities

Traders Union
All News
Financial Institutions
IMF on AI-driven payments challenges

Highlights

IMF official Dan Katz stated that AI agents can enhance payment system efficiency by reducing switching costs and automating complex transactions, especially in $1 quadrillion annual cross-border payments.
Katz highlighted initiatives like CHIPS tokenized deposit settlement, ECB's Project Pontes, and Bank of England stablecoin reserve requirements as key examples for programmable, interoperable payments infrastructure.
IMF emphasizes the need for updated regulatory frameworks to address legal status of tokens, rising liquidity needs from atomic settlement, and heightened systemic cyber risks in tokenized markets.

AI applications and payment system redesign

As reported by IMF, citing remarks by IMF official Dan Katz at a SWIFT conference, the Fund sees AI as a tool that can improve payments even before deeper changes to financial market infrastructure are completed. Katz said AI agents can reduce information frictions by comparing fees, exchange rates and execution quality in real time, potentially lowering switching costs and pushing providers to improve pricing and services.

He also said AI can automate complex transactions that still require significant human input, including machine-to-machine payments and decisions over timing based on liquidity, costs and contractual obligations. In cross-border payments, where legal systems, currencies and compliance requirements add extra complexity, AI agents could also speed up due diligence, sanctions screening and other compliance functions.

Katz argued that many benefits often associated with a redesigned financial system could be achieved by upgrading existing payment architecture. He said IMF staff estimate that cross-border payments move nearly $1 quadrillion a year, meaning even small efficiency gains could produce significant economic value.

Beyond near-term improvements, Katz said the financial system will need to become more digital, interoperable and programmable if AI agents take on a larger role in economic activity. He pointed to tokenization as a potentially important technology because tokenized assets can be programmed and settled almost instantly, and cited initiatives including a CHIPS multi-bank tokenized deposit settlement platform, the ECB's Project Pontes and the Bank of England's stablecoin reserve requirements.

Regulatory choices and systemic risks

Katz said policymakers need to support innovation without weakening stability, which requires legal and regulatory frameworks that allow payment technologies to compete on their merits. He said authorities need to clarify the legal status of tokens, define their relationship to real-world assets and avoid unnecessary barriers to interoperability across networks.

He also said liquidity rules will need to keep pace with new payment models because atomic settlement may reduce counterparty risk while increasing intraday liquidity needs through prefunding. The speech added that stablecoins and tokenized bank liabilities raise broader questions about whether central banks and finance ministries should provide new forms of tokenized public money or government obligations.

On risk, Katz said cyber threats remain the most immediate concern as AI increases the speed and scale of vulnerability discovery and exploitation. He warned that tokenized markets rely heavily on shared infrastructure and third-party providers, which could allow a cyber incident to spread widely enough to become systemic.

The IMF's policy approach, he said, is to identify clear market failures and assess whether current regulation is sufficient before introducing new rules. Where fresh intervention is needed, Katz said it should be proportionate, targeted and internationally coordinated so that competition and innovation are not unnecessarily constrained.

Agentic AI tools that automatically optimize household cash returns could amplify deposit outflows from traditional banks, our publication previously noted. The earlier piece explained that if consumers use AI assistants to sweep idle balances into higher-yield fintech or brokerage alternatives, banks could lose a key source of low-cost funding, adding pressure to the sector already facing higher rates.

This material may contain third-party opinions, none of the data and information on this webpage constitutes investment advice according to our Disclaimer. While we adhere to strict Editorial Integrity, this post may contain references to products from our partners.