AI Incident Response and Business Continuity CRAGE Practice Questions (Page 9)

Questions 41–45

1. 41expert · hard

   An organization has conducted a tabletop exercise for its AI incident response plan and identified several gaps. What is the MOST effective way to address these gaps?

   • AIgnore the gaps because tabletop exercises are not realistic.
   • BReplace the incident response team with a new team that can handle the gaps.
   • CDocument the gaps and move on to the next exercise.
   • DUpdate the incident response plan to address the gaps and schedule a follow-up exercise to validate the changes.
   Reveal Solution Discussion Select an answer first
2. 42application · easy

   During an AI incident, the response team has detected unusual model behavior and is trying to determine the root cause. They have collected logs and are analyzing them. Which phase of the AI incident response lifecycle are they in?

   • ALessons learned
   • BRecovery and post-incident review
   • CDetection and triage
   • DContainment and eradication
   Reveal Solution Discussion Select an answer first
4. 43foundation · easy

   What is the primary purpose of redundancy in an AI system?

   • ATo simplify the model architecture
   • BTo ensure high availability by having backup components
   • CTo improve the accuracy of the model
   • DTo reduce the amount of training data needed
   Reveal Solution Discussion Select an answer first
5. 44expert · hard

   A company has a mature IT incident management process but is new to AI incident response. They have just experienced an AI incident where the model produced biased outputs. The IT team followed the standard IT incident process, but the AI-specific steps were missed. What should the company do to improve integration?

   • AReplace the IT incident management process with a new AI-specific process.
   • BRely on the IT team to learn from this incident without updating any documentation.
   • CUpdate the IT incident management playbooks to include AI-specific response steps and train the team on them.
   • DCreate a separate AI incident response team that handles all AI incidents independently.
   Reveal Solution Discussion Select an answer first
6. 45application · medium

   A healthcare AI system that predicts patient readmission has been compromised by a prompt injection attack that alters model outputs. The security team has isolated the affected model version. What should the team do NEXT to contain and eradicate the incident?

   • ARoll back to the last known-good model version and quarantine the compromised model's training data.
   • BDelete the compromised model and all its logs to prevent further spread.
   • CImmediately restore service using the same model but with additional monitoring.
   • DShut down the entire AI system and do not restart until a full investigation is complete.
   Reveal Solution Discussion Select an answer first