Autonomous AI agents require new payment safeguards, Waller warns
Federal Reserve Governor Christopher Waller outlined the structural impact of autonomous AI agents on payment systems. Speaking at an industry event on October 5, 2026, Waller highlighted the need for new authentication protocols, liability frameworks and recalibrated fraud models.
Delegated shopping tests current rails
Waller distinguishes between agent-assisted models, where humans retain execution control, and agent-delegated models, where AI agents receive autonomous spending authority with preset guardrails.
Consumer transactions will pioneer this shift, but business-to-business commerce offers significant scale due to recurring purchasing rules and supplier constraints.
However, higher transaction values in corporate procurement magnify the financial exposure from unauthorized actions or algorithmic errors.
Autonomous agents will also rely heavily on continuous machine-to-machine micropayments for data queries and computational tasks, favoring payment networks with lower flat transaction fees across ACH, wire, instant rails and payment cards.
The asymmetry of automated defense
Scaling agentic commerce hinges on resolving critical governance gaps across authentication, liability and fraud detection.
Proving an agent possesses valid authority shifts the traditional payer verification model, while unclear fault allocation complicates mistaken purchases.
In cross-border flows, large language models reduce false-positive screening alerts, yet cybersecurity risks grow as attackers target single vulnerabilities across an expanding defensive surface.
Sound warnings before the wave hits
Waller correctly identifies the legal void in agent liability before commercial rollout accelerates.
Yet central banks remain passive while private networks dictate the technical architecture.
Without swift regulatory standards, payment providers will carry unmanageable risks from automated errors.
Source: Payments in the age of AI agents
IN: