フリー問題

ISACA Advanced in AI Risk のフリー問題 6 / 20 問目

問題文

A credit provider deploys a model that recommends whether to decline an application. Regulators expect meaningful human oversight. Which arrangement best satisfies that expectation?

選択肢

  1. A sample of five percent of declined applications is reviewed each month by a second-line analyst who reports aggregate agreement rates to the risk committee.
  2. Every declined application is reviewed by a human within one business day and the reviewer records whether they agree or disagree with the recommendation before the decision is communicated to the applicant.
  3. The model is configured to abstain when its confidence is low and abstentions are routed to a human underwriter for a full manual assessment of the application.
  4. Reviewers receive the reasons behind each recommendation and have authority to overturn it and are measured on decision quality rather than throughput.

解答・解説を確認するには

正解と解説の確認、回答の記録には無料登録が必要です。登録すると演習モードでフリー問題に回答し、正誤と解説をその場で確認できます。