CCAR-P: Claude Certified Architect — Professional

Question

An architect is assessing a Claude-based candidate-screening tool before deployment. A stakeholder claims that no bias evaluation is needed because the model was not trained on company data.

Which two responses most accurately dispute this claim? (Select two.)

Choose two

Explanation

A model can retain demographic biases from its broad pretraining even when it has not been trained on an organization’s data. For an employment-screening use case, fairness evaluation should measure whether outcomes differ across protected groups before deployment. NIST guidance calls for evaluating fairness and bias and using context-specific metrics to assess performance and harms across groups.

Learn more

CCAR-P: Claude Certified Architect — Professional

AstroTutor

Notes

Readiness

QuestionQ23

Community Discussion