Model risk
The potential for adverse consequences from decisions based on model outputs, including misuse of a sound model.
US banking agencies' revised model risk guidance, issued as SR 26-2, describes model risk as the potential for adverse financial consequences associated with models, which may result from decisions made based on model output, and notes that even a sound model can carry high risk if it is misapplied or misused. The guidance states that generative and agentic AI models are outside its scope, while its principles apply to traditional models and non-generative, non-agentic AI. Banks are still expected to govern agent tools under their broader risk management, and the UK's PRA sets comparable model risk principles in SS1/23.
Agent Minute explains this term on 12 October 2026.