
New governance frameworks emerge for generative AI in financial institutions
A July 2026 research paper proposes a Generative AI Control Framework compatible with U.S. supervisory guidance such as SR 26-2, outlining governance, risk control, and monitoring practices specific to generative AI workflows in financial institutions. In parallel, other work this year integrates OCC, SR 11‑7, CFPB, and FinCEN requirements into a lifecycle framework for AI-driven fraud detection models.
There is an opportunity to move faster on front‑office generative AI—such as advisory copilots, document automation, and customer support—by adopting emerging control frameworks that map directly to existing model risk and compliance expectations, making it easier to secure board and regulator approval.
highDeploying generative AI at scale without a framework aligned to supervisory guidance heightens the risk of model risk findings, mis-selling, biased outcomes, and enforcement, especially as regulators scrutinize AI in fraud, credit, and customer interactions.
high