Control #

E

4

.

1

Evaluate AI outcomes across relevant groups

Regularly test AI outcomes for disparities across protected classes and other relevant groups. Document findings and track changes over time.

Evidence

Bias audit logs that show periodic compliance checks across deployments

Documentation of bias and protected class testing methodology

Evidence of bias and protected class simulations and testing

Recommended actions

We'll recommend specific practices and actions for complying with this control.

Provide feedback on this control