Glossary
Model risk
The risk of loss from decisions based on incorrect or misused models.
Model risk is the potential for adverse outcomes (financial loss, poor decisions, or reputational and regulatory harm) arising from errors in a model or from using a model incorrectly. Managing it is the subject of SR 26-2.
Generative AI introduces model risk in new forms: non-determinism, prompt sensitivity, and shadow usage. Bringing LLMs into a model inventory with monitoring and documentation is how firms keep that risk governed.