Exhibit B
The risk-assessment section of the NAIC AI Evaluation Tool, covering identification, measurement, and mitigation of AI risks including unfair discrimination.
Exhibit B of the NAIC AI Evaluation Tool focuses on risk assessment. It asks insurers to identify the AI systems used in their insurance practices, assess the risks those systems pose, and document the controls used to mitigate them.
Important concepts in Exhibit B include:
- Mapping AI systems to specific insurance practices, such as underwriting, pricing, claims, or marketing.
- Risk tiering based on the potential for adverse outcomes, including unfair discrimination or operational failure.
- Testing to validate that AI systems perform as intended, both before deployment and on an ongoing basis.
- Remediation when risks are identified.
Exhibit B is where most of the model-risk-management work appears. Insurers typically need model inventories, risk ratings, validation reports, and monitoring plans. Our NAIC AI Evaluation Tool guide covers the evidence examiners expect to see.