What's Happening?
On November 4, 2026, the Office of the Comptroller of the Currency (OCC) released Bulletin 2026-54, titled 'Super Intelligence: Supervisory Framework.' This framework outlines how OCC examiners will evaluate banks' use of artificial intelligence (AI)
in various ordinary business activities, including credit decisioning, payments, fraud detection, customer service, cybersecurity, and risk management. The OCC, consistent with the White House's preferred terminology, refers to AI as 'super intelligence' (SI) throughout the framework, primarily focusing on generative and agentic AI systems. Generative AI uses neural networks to analyze data and produce unique outputs, while agentic AI systems can independently plan, execute tasks, make decisions, and interact with other systems with limited human intervention. The framework applies a technology-neutral, risk-based approach, recognizing that AI can introduce, amplify, or modify existing operational, compliance, credit, and other banking risks. It does not establish new supervisory expectations or enforceable standards but provides insight into areas the OCC may review during examinations, such as agent permissions, testing, auditability, operational resilience, and third-party vendor arrangements.
Why It's Important?
This new supervisory framework from the OCC is important because it provides a clear roadmap for how U.S. banks and their technology providers should approach AI governance and risk management. By emphasizing a risk-based approach, the OCC acknowledges that not all AI applications carry the same level of risk, and controls should be proportionate to the potential consequences of an AI system's failure or misuse. This guidance is crucial for maintaining the safety and soundness of the federal banking system as financial institutions increasingly adopt AI technologies. The framework's focus on agentic AI, which can take autonomous actions, highlights growing concerns about the potential for significant financial or customer consequences if these systems are not properly managed. For banks, understanding and implementing these guidelines will be essential to avoid regulatory scrutiny and potential penalties, while for technology providers, it will shape the development and deployment of AI solutions for the financial sector. The framework also underscores the importance of robust third-party risk management, as many banks rely on external vendors for AI solutions, impacting contractual liability and incident notification protocols.
What's Next?
Banks are expected to align their AI governance processes with the OCC's new framework, particularly by classifying AI use cases according to the illustrative four-tier methodology. This will involve evaluating whether their current AI inventories and governance frameworks adequately address the risks associated with different levels of AI autonomy and customer impact. Financial institutions will need to review and potentially revise their vendor agreements to ensure they reflect the appropriate risk tier of the services provided, incorporating stronger contractual protections for higher-tier deployments. Key areas of focus will include clearly defining the scope of an agent's capabilities, establishing human oversight mechanisms (human-in-the-loop or human-on-the-loop), and enhancing monitoring, testing, and auditability of AI systems. The OCC's emphasis on operational resilience and intervention means banks should develop or refine AI-specific incident response mechanisms and ensure technical and contractual arrangements support timely intervention when autonomous systems behave unexpectedly. Upcoming renewals and new procurements of AI technologies will offer timely opportunities for banks to integrate these protections into their agreements.
Beyond the Headlines
The OCC's 'Super Intelligence' framework signals a significant shift in regulatory thinking, moving beyond treating AI as a distinct risk category to integrating it within existing operational, compliance, and credit risk frameworks. This approach acknowledges the pervasive nature of AI and its potential to amplify or modify traditional banking risks. The distinction between informational ('read') access and transactional ('write') access for AI agents highlights a critical ethical and legal dimension: the degree of autonomy granted to AI and the corresponding accountability. As AI systems become more autonomous, questions of liability for errors, customer losses, and unauthorized actions will become increasingly complex. The framework's emphasis on human oversight, whether 'in-the-loop' or 'on-the-loop,' underscores a broader societal debate about the balance between technological advancement and human control, particularly in sectors with significant public impact like finance. This guidance could also influence future regulatory approaches to AI in other critical U.S. industries, setting a precedent for how government bodies will manage the risks and opportunities presented by rapidly evolving AI technologies.













