A retail company builds a resume-screening model using five years of past hiring decisions as training data. Most of those past hires came from a small set of universities that historically enrolled few students from certain underrepresented backgrounds. After deployment, the model consistently scores equally qualified candidates from those backgrounds lower than candidates from the well-represented universities. Which responsible AI principle is most directly at risk in this scenario?

The responsible AI principle most directly at risk is fairness. The model is producing systematically unequal outcomes for equally qualified candidates based on a characteristic tied to their background, which is precisely the kind of biased treatment the fairness principle is meant to catch and prevent. The reliability and safety option is incorrect because reliability and safety concerns whether a system performs consistently and safely under the conditions it was designed for, such as avoiding crashes or unsafe recommendations, not whether its scores are skewed across groups. The transparency option is incorrect because transparency is about disclosing how a system works, what data it uses, and what its limitations are so users can make informed judgments; the scenario does not describe a disclosure problem, it describes an outcome problem. The accountability option is incorrect because accountability concerns having identifiable people or processes responsible for a system's outcomes and able to intervene when something goes wrong; while accountability structures might eventually help catch this issue, the scenario itself describes the biased scoring pattern, which is a fairness violation at its core.