Investigate an AI response leak across input, retrieval, and policy changes.
Correlate multiple service logs.
Choose two audit practices.
Use common request ID, principal, time, model, document, and policy versions, with synchronized time, retention, access separation, and tamper detection.
Detailed explanation
Processing and changes can be correlated.
Processing and changes can be correlated.
Evidence is reliable and searchable.
Evidence is reliable and searchable.
Correlation becomes uncertain.
Correlation becomes uncertain.
Sensitive exposure expands.
Sensitive exposure expands.
Try it yourself
An example you can run in a temporary verification environment.
AWS CloudTrail・CloudWatch Logs公式とAIF-C01 Domain 5.2の監査相関を確認する。Expected result
複数サービスのAIイベントを相関し、証跡と機密性を両立できる。Key points
- Correlation ID
- Time sync
- Tamper detection
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.