IRS Risk Chief Warns Automated Finance Requires AI-Driven Regulatory Supervision

Internal Revenue Service (IRS) Chief Risk and Control Officer Dottie Romo warned that traditional regulatory oversight is too slow to monitor autonomous financial systems, arguing that regulators may need to deploy algorithmic supervision to keep pace with machine-driven payments. Speaking on a panel hosted by BeInCrypto at the Future of Money summit at United Nations Headquarters in New York, Romo was joined by Dino Cataldo Dell'Accio of the UN Joint Staff Pension Fund and Julius Moye, Manager at Mastercard's Financial Crime Solutions, to address the rapid rise of autonomous AI agents. ## The Scale of Autonomous AI Agent Payments Recent research titled State of AI Agent Payments 2026 highlights the growing volume of machine-driven commerce on blockchain rails. Between July 23 and August 26, researchers tracked 6.4 million x402 payment transactions totaling $119,947 across the Base and Solana networks. Although 90.8% of these micro-transfers were valued at less than one cent—paying for automated API access, data, and compute resources—the x402 protocol has recorded nearly 200 million settlement transactions since launching. ## Algorithmic Supervision and Systemic Safeguards Romo emphasized that relying on periodic risk reports is insufficient when automated agents execute millions of financial decisions per minute, advocating for real-time monitoring while keeping human oversight for key decisions. Highlighting past market crashes such as the 2012 Knight Capital trading disaster and the Terra/Luna collapse, Mastercard's Julius Moye proposed a hybrid control model where AI systems automatically contain operational anomalies before escalating critical issues to human managers. Meanwhile, Dell'Accio stated that accountability must trace back through three core questions: who developed the code, who implemented it, and who oversees it. The panel follows broader US oversight scrutiny, including recent actions where FTC regulators opened probes into AI developers and federal agencies updated unhosted wallet rules. Key Takeaways: - IRS Risk Chief Dottie Romo called for real-time automated supervision to track machine-driven finance. - 6.4 million x402 transactions totaling $119,947 occurred on Base and Solana between July 23 and August 26, with 90.8% under $0.01. - The x402 agent ecosystem has logged nearly 200 million settlement transactions since its launch. - Mastercard's Julius Moye cited the 2012 Knight Capital disaster and Terra/Luna collapse when advocating for AI automated containment safeguards. ## Why It Matters As machine-to-machine microtransactions proliferate across public blockchains, financial oversight built around manual reporting and delayed audits becomes obsolete. Establishing automated algorithmic monitoring is essential to prevent systemic liquidity shocks or rogue agent behavior from destabilizing broader markets. Moving forward, market participants must watch whether regulatory bodies establish formal frameworks for AI containment mechanisms and machine-readable compliance protocols.