Câu 10: 612-51: Certified Responsible AI Governance and Ethics Professional

A global cybersecurity consulting firm is challenged by its board after noticing unexpected patterns in the AI-driven candidate ranking system used for selecting red-team operators. Despite a diverse applicant pool, the model persistently scores candidates with certain certification keywords significantly higher than equally qualified individuals with robust exploit-development portfolios, active bug-bounty histories, or advanced defensive threat-hunting experience. Internal analysis confirms that while the dataset was balanced, the model learned to overweight specific textual features during training, leading to disproportionate ranking outcomes. As the Head of Responsible AI Operations, you must determine the fundamental bias source to guide remediation and ensure fair evaluation standards across global hiring pipelines. Which type of bias is being exhibited by the AI system in this scenario?

1. A. Data Bias
2. B. Confirmation Bias
3. C. Societal Bias
4. D. Algorithmic Bias — đáp án hiện tại