Supporting case
The OCC tells bank examiners to assess AI by its authority and consequences
What happened
On November 4, the Office of the Comptroller of the Currency issued a supervisory framework for examiners reviewing covered banks’ use of what the bulletin calls “super intelligence,” or SI. Examiners should tailor their work to a bank’s size, complexity and use cases, with attention to consequences, autonomy and human oversight. A simple system allowed to initiate transactions can require closer controls than a more sophisticated system used only for internal assistance.
The bulletin expressly says the guide creates no new supervisory expectations for banks, enforceable standards or prescriptive requirements. Banks need not use it to assess their systems, create a separate SI risk category or define a separate risk appetite. It supplements examination of existing governance, compliance, technology and operational risks. Official release ↗; corroborating reporting ↗.
This is material institutional practice: it tells supervisors to look at delegated authority and practical harm, rather than model sophistication alone. But it does not establish an additional right or remedy comparable to ASIC’s adopted trading obligations. Our authoritative-existing-law criterion permits scoring a consequential clarification before enforcement; it does not make every examiner framework such a clarification.
Dates and implementation stage
Action/stage 2026-11-04; reporting 2026-11-04; approved 2026-11-05
