Question

A company builds a generative AI tool to draft job candidate evaluation summaries. During testing, leadership notices the tool consistently uses more favorable language for candidates from certain demographic groups. An investigation reveals the model was trained on years of the company's historical hiring notes. Which responsible AI concern does this situation most directly illustrate?

Accepted Answer

Bias inherited from historical training data that can produce unfair outcomes for certain groups

Answer

Hallucination, because the model is fabricating candidate qualifications that do not exist

Answer

A context window limitation preventing the model from processing all candidate data

Answer

A privacy breach caused by exposing candidate personal information to external parties