Separating system instructions and user input | AWS Certified AI Practitioner (AIF-C01) original practice question

2.3Standard

Insert a user request into an internal FAQ prompt.

Prevent the request from changing safety rules or permissions.

Choose two appropriate implementations.

Hint 1: Focus on the requirement

Identify what must change and what must remain unchanged. Then choose the command whose scope matches that requirement.

Hint 2: Check each command's role

Separate commands that display information, change state, remove data, or affect a child process.

信頼された指示と未信頼入力を別の役割・境界として扱う: Untrusted text must not override system policy.
認可や危険操作の判定をモデルの文章だけに委ねない: A trusted server-side control must enforce permissions.
ユーザー入力へ管理者権限を付ける: Text input has no authority.
指示を隠せば入力検証は不要になる: Hidden prompts do not stop malicious input or output.

Detailed explanation

CorrectKeep trusted instructions and user text in separate roles or boundaries

Untrusted text must not override system policy.

CorrectDo not delegate authorization or dangerous actions to model prose

A trusted server-side control must enforce permissions.

IncorrectGive user text administrator authority

Text input has no authority.

IncorrectHidden instructions remove the need for input validation

Hidden prompts do not stop malicious input or output.

An example you can run in a temporary verification environment.

AWS公式生成AIセキュリティとAIF-C01 Domain 2.3のプロンプトインジェクション対策を確認する。

役割分離とサーバー側認可の必要性を説明できる。

Foundation review

Check whether the command acts on the current shell, a new process, an existing process, or a file.

Use the supplied command and compare the output with the expected result.

Question IDAWS-AIF-117

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