A public chatbot receives requests for secrets and harmful content.
Control both sides of the model boundary.
Choose two appropriate controls.
Inspect input for prohibited content and injection indicators, and inspect output for secrets, harmful content, and policy violations.
Detailed explanation
Dangerous requests can be rejected or isolated before generation.
Dangerous requests can be rejected or isolated before generation.
Input checks alone cannot guarantee safe output.
Input checks alone cannot guarantee safe output.
A model can still generate unsafe content.
A model can still generate unsafe content.
Logs need minimization, access control, and retention rules.
Logs need minimization, access control, and retention rules.
Try it yourself
An example you can run in a temporary verification environment.
Amazon Bedrock Guardrails公式資料とAIF-C01 Domain 5.1の入出力保護を確認する。Expected result
入力だけでなく出力も検査する多層防御を説明できる。Key points
- Input
- Output
- Defense in depth
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
- Command output formatting can vary slightly by distribution or tool version.
- Run the example in a temporary directory or process when possible.
Foundation review
Read the scope first
Check whether the command acts on the current shell, a new process, an existing process, or a file.
Verify the observable result
Use the supplied command and compare the output with the expected result.