5.1Standard
The system must be tested for injection and harmful output.
Make attacks reproducible.
Choose two practices.
× Incorrect
Define scope, success criteria, and prohibited actions, test multilingual and multi-turn boundaries, and record model, attack, mitigation, and residual risk.
Detailed explanation
CorrectDefine scope, prohibited actions, success criteria, languages, turns, and boundaries
Test coverage is managed.
IncorrectTest dangerous actions on production users without authorization
Impact and permissions are unsafe.
CorrectRecord input, model version, output, detection, mitigation, and residual risk and retest
Fix effectiveness is traceable.
IncorrectReport only successful examples and hide attacks
Risk and reproducibility are lost.
Check official documentation
Compare the following points with the official documentation.
AWS公式AIF-C01 Domain 4のAI安全試験、レッドチーム、残余リスクを確認する。
Official references
What to confirm
攻撃例を再現可能な試験ケースへし、緩和策の効果と残余リスクを報告できる。
Key points
- Red team
- Residual risk
- Retest
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
Red team
An authorized group that probes a system for security or safety weaknesses.
Attack surface
The inputs, interfaces, tools, and behaviors that can be targeted.
Question IDAWS-AIF-466
EnvironmentAWS公式AIF-C01試験ガイドとAWS公式ドキュメントの確認
Last technical review2026-08-20
Report an error or rights issue