A hiring tool uses Claude to score candidate resumes on a 1-10 scale for job fit. After 3 months in production, an HR audit discovers that candidates with names typical of certain ethnic groups receive systematically lower scores on equivalent qualifications. What is the immediate and long-term architectural response?
Select an answer to reveal the explanation.
Short Explanation and Infographic
Here's the deal — b is correct because discovering systematic demographic bias in a hiring tool requires immediate mitigation (suspend the tool, conduct audit, notify affected parties) and long-term architectural remediation (anonymization, bias testing in prompts and training data, ongoing monitoring). This is both an ethical requirement and a legal one — employment discrimination laws apply to AI-assisted hiring decisions.
Full explanation below image
Full Explanation
B is correct because discovering systematic demographic bias in a hiring tool requires immediate mitigation (suspend the tool, conduct audit, notify affected parties) and long-term architectural remediation (anonymization, bias testing in prompts and training data, ongoing monitoring). This is both an ethical requirement and a legal one — employment discrimination laws apply to AI-assisted hiring decisions. A is wrong because adding a disclaimer while continuing to use the biased tool does not mitigate harm — biased decisions continue to be made, and a disclaimer does not remove liability for disparate impact discrimination. C is wrong because re-prompting alone is insufficient — names may have been irrelevant from the start but bias may have been encoded through correlated signals in resume text (geography, school names, extracurriculars); a name-ignoring prompt does not address these proxy variables. D is wrong because retraining on more diverse data is a long-term mitigation that takes significant time — it does not address the immediate harm to candidates already scored with the biased system.