AI トレーニング データのバイアス テストが完了したら、テストの有効性を検証するために最も重要なのは次のどれですか。
正解:B
Even after identifying and mitigating bias, organizations must ensure that AI outputs do not create unacceptable risks. AAIA emphasizes that bias mitigation must result in: Fair outcomes Justifiable predictions No disproportionate harm to any demographic group Alignment with organizational risk tolerance Option B reflects this requirement, ensuring that the model's real-world impact aligns with documented risk thresholds.