A compliance officer discovers that the organization's AI hiring tool was trained on application data where the offer/no-offer label was determined by the previous hiring manager, who had documented complaints of gender discrimination. Which data integrity issue is MOST critical?
Select an answer to reveal the explanation.
Short Explanation
Here's the deal — you've got label bias because the offer/no-offer labels came from a hiring manager with documented gender discrimination complaints. Those biased decisions become the model's ground truth, so the AI learns to copy the manager's discriminatory pattern regardless of the features used.
Full Explanation
The most critical issue is label bias, also called ground truth contamination. The offer/no-offer labels were created by a previous hiring manager with documented gender discrimination complaints, so the labels encode biased historical decisions. The model learns to reproduce that pattern, even if the features are appropriate. The temporal bias option about data predating current hiring best practices concerns staleness, not contaminated outcomes. The sampling bias option about only applicants passing the initial resume screen concerns which examples were included. The feature engineering bias option about wrong features would involve different features driving different problems, not the corrupted ground truth.