A hiring-support model has a higher error rate for one group.
Do not deploy it unchanged; investigate the impact and improve the system.
Which responsible-AI activity is appropriate?
Measure group-level performance, investigate data and model causes, improve the system, and include human review for high-impact decisions.
Detailed explanation
This reduces transparency and prevents appropriate recourse.
High-impact use requires risk-based review and improvement.
An average can hide unequal impacts on specific groups.
Responsible AI combines measurement, diagnosis, mitigation, and operational oversight.
Check official documentation
Compare the following points with the official documentation.
AWS公式AIF-C01 Domain 4のResponsible AIと公平性に関する説明を確認する。
Official references
What to confirm
公平性をグループ別に評価し、人間のレビューや改善サイクルを組み込む理由を説明できる。
Key points
- Fairness
- Group-level evaluation
- Human review
Notes
- Environment: AWS公式AIF-C01試験ガイドとAWS公式ドキュメントの確認
- Compare the definition with the AIF-C01 objective and the service documentation before applying it.
- Keep the concept separate from a particular answer choice or implementation step.
Foundation review
Average limitation
A strong overall score can hide harmful error differences between groups.
Mitigation loop
After a mitigation, remeasure group outcomes and business impact to check that the change helped.