The Governance Gap in Financial Markets: Risks of Agentic AI

On April 17, 2026, the Federal Reserve issued SR 26-2, a revised guidance on model risk management, which notably excluded generative and agentic AI from its scope. This exclusion raises concerns about the regulatory framework governing the rapid adoption of these technologies in banking, particularly as institutions like JPMorgan Chase and Goldman Sachs increasingly deploy agentic AI systems for critical operations.

Forbes

The adoption of agentic AI in financial services is accelerating. JPMorgan Chase has allocated approximately $18 billion annually to technology, resulting in over 400 production use cases by early 2026. Similarly, Lloyds Banking Group anticipates that agentic AI will generate £100 million in value this year by automating tasks such as fraud investigations. Other major banks, including Citigroup, Morgan Stanley, and Bank of America, are embedding AI across various operations, transitioning from AI as an assistant to AI with transactional authority.

Forbes

The deployment of agentic AI introduces significant risks, including data leakage, unauthorized actions, and algorithmic bias. Deloitte's analysis identifies over 350 distinct risks associated with autonomous behavior, many of which could threaten banking systems. A critical concern is the potential for coordinated actions among multiple banks' AI agents, which could lead to market volatility or liquidity crises. The 2010 flash crash, which erased approximately $1 trillion in market value in minutes, serves as a historical precedent for such risks, highlighting the dangers of interconnected algorithms.

Forbes

The Bank of England's Financial Policy Committee has acknowledged that while advanced AI is not currently posing systemic risks in UK finance, the situation could change rapidly as firms adopt agentic AI. The Federal Reserve's decision to exclude agentic AI from SR 26-2 reflects a recognition of the technology's rapid evolution, but it also leaves a significant governance gap. Current regulatory frameworks are designed for static algorithms and do not account for the dynamic nature of agentic systems.

Forbes

To address the governance gap, a new framework is necessary. This framework should include continuous validation of AI systems, interaction testing to assess systemic risks from multiple agents, and concentration monitoring to identify correlated behaviors among institutions using similar underlying models. The Bank of England's scenario analysis is a step in this direction, but the U.S. lacks a comparable program.

Forbes

Critics argue that the current regulatory landscape is ill-equipped to manage the complexities introduced by agentic AI. Risk officers within banks have expressed concerns that existing validation processes are inadequate for assessing autonomous systems that interact with live markets. The fragmented efforts to establish internal AI governance functions further complicate the regulatory environment.

Forbes

The rapid deployment of agentic AI in banking presents unprecedented challenges for financial regulation. As institutions increasingly rely on these technologies, the absence of a robust governance framework poses significant risks to market stability. Without timely intervention to close the regulatory gap, the financial sector may face consequences reminiscent of past market disruptions.

Forbes

• “The agents being deployed today are more autonomous, more interconnected, and less predictable than the algorithms that caused that crash.” — Zennon Kapron, Financial Analyst

Forbes

• “It is about the absence of supervisory tools designed for systems that learn, adapt, and interact.” — Zennon Kapron, Financial Analyst

Forbes

• “The risk compounds because these agents share underlying model architectures.” — Zennon Kapron, Financial Analyst

Forbes

• “The regulators know.” — Zennon Kapron, Financial Analyst

Forbes