Skip to content
AI Security

Interview question · AI Red Teaming, Evaluation & Response

How do you preserve evidence in an AI incident when the logs contain sensitive prompts and personal data?

Preserve relevant evidence within its authorization boundary. Protect collection, record gaps and transformations, and explain what a checksum does and does not establish.

hardNewUpdated Nov 2026

Preserve relevant evidence within its authorization boundary. Protect collection, record gaps and transformations, and explain what a checksum does and does not establish.

Make your first attempt

What would you say before reading the answer?

  1. 01 State the assumptions your answer depends on.
  2. 02 Explain the mechanism and choose a control.
  3. 03 Name a condition that would change your decision.

The worked answer includes an approach, scoring criteria and follow-up questions. Compare the reasoning with your own.

10 answers per topic instead of 5, plus saved progress and bookmarks · costs nothingor unlock all 529 remaining answers · $25

Every worked answer in all 17 topics, every lesson in every course, the full concept curriculum, the complete practice-test bank, and the downloadable security field guides.

6 months · One payment · No auto-renewal

What if · the follow-ups in this answer

The answer works through 4 follow-ups like an interviewer would

In an original fictional incident, a connector returns another customer’s record. The trace contains personal information and a credential-shaped value, and routine retention will remove it soon.

  1. What if the complete trace contains data the team is not authorized to retain?
  2. What if the collected archive’s checksum matches tomorrow?
  3. What if the customer requests the complete archive?
  4. What would you hand to the investigation lead?

Each rung has a nudge, a model answer, the principle that carries over and the weak answer to avoid. Premium opens every answer and ladder.

UP NEXT ON YOUR JOURNEY
Next in this trackHow would you detect someone stealing your model through its API?Next in this trackDesign LLM-assisted phishing triage for a security team, and list its failure modes.Next in this trackDesign an automated red-teaming agent that finds new failure classes in a target system from its policy.
DISCUSSION · 0

No comments yet. Share a correction or a different way to answer it.