OCC Issues Supervisory Framework for Bank Use of AI and Agentic AI
OCC Bulletin 2026-54 gives US bank examiners a framework for assessing AI use, based on consequences rather than technology sophistication, with three key principles on scope, risk categorization and examination intensity.
The US Office of the Comptroller of the Currency (OCC) has issued Bulletin 2026-54, titled 'Super Intelligence: Supervisory Framework', providing examiners with a framework for assessing banks' use of AI, including agentic AI, across activities such as credit decisioning, payments, fraud detection, customer service, operations and risk management.
The bulletin sets out three key principles. It is an examiner tool rather than a rule and does not establish new supervisory expectations or enforceable standards. AI is treated not as a separate risk category but as a factor that amplifies or modifies existing risk areas such as operational, compliance and credit risk. Examination intensity is determined by consequences rather than technology sophistication, scaling with the potential impact of failure, the degree of decision-making transferred from humans to systems, and the level of human oversight.
The framework applies to US national banks and does not constitute a legal obligation outside that jurisdiction. This summary is based on secondary legal analysis of the bulletin.