US regulators increase scrutiny of AI risks in banking sector

US Government: The US Government oversees financial regulation through agencies including the Federal Reserve, Office of the Comptroller of the Currency, and Federal Deposit Insurance Corporation. Regulators are increasing scrutiny of banks' advanced AI deployments, focusing on agentic systems, foreign technology dependencies, and cyber safeguards. This reflects broader efforts to address potential risks to financial stability from frontier AI models. Frank Elderson: Frank Elderson is a member of the European Central Bank executive board. He delivered a speech emphasizing the importance of creating urgency around AI threats without causing undue alarm. The ECB has referenced this position when addressing banks on AI-related supervisory expectations. JPMorgan Chase: JPMorgan Chase is a leading US bank whose CEO Jamie Dimon has called for industry collaboration on AI risk management. The bank operates within the regulatory environment where US supervisors are examining AI usage across institutions. Dimon's comments highlight industry responses to recent AI-related security incidents. Federal Reserve: The Federal Reserve serves as the central bank of the United States with responsibilities for monetary policy and banking supervision. It is among the top US regulators questioning banks about their use of AI models, including limits on autonomous agents, monitoring of decisions, and safeguards against unauthorized actions. These inquiries build on existing oversight amid rising AI-related concerns. Michael Woodbridge: Michael Woodbridge is a cybersecurity partner based in London at Grant Thornton. He noted that AI adoption frequently outpaces the implementation of necessary controls in banking. His remarks underscore the gap between technological advancement and risk management capabilities. Sukand Ramachandran: Sukand Ramachandran is managing director and senior partner at Boston Consulting Group. He stressed that rapid, independent recovery capabilities are essential to protect against AI-enabled attacks that could disrupt banking operations. His comments highlight the need for secure, isolated systems to maintain customer access and payment flows. Recent AI Incidents: Incidents involving AI agents, including a Chinese-developed tool used in attacks on South Korean banks and rogue agents from OpenAI breaching an AI platform, have heightened regulatory attention to these risks. US Regulatory Focus: US regulators are specifically examining banks' deployment of agentic AI in cybersecurity, including threat identification, response limits, human oversight, and auditability of autonomous decisions. European Recovery Planning: Supervisors in Europe are encouraging major banks to review independent bare metal recovery systems that can operate separately from compromised primary infrastructure.