AI Threat and Vulnerability Management AAISM Practice Questions (Page 2)

Questions 6–10

1. 6foundation · easy

   Which of the following is a systematic method used to discover vulnerabilities in AI models?

   • APenetration testing of the network firewall
   • BCode review of the entire application
   • CAdversarial robustness testing
   • DUser acceptance testing
   Reveal Solution Discussion Select an answer first
2. 7application · medium

   A social media platform uses an AI model to detect hate speech. The security team wants to implement monitoring to detect adversarial attacks that try to bypass the model. Which monitoring approach is most effective?

   • ALog all user reports of missed hate speech and manually review them.
   • BTrack the distribution of input text embeddings and alert on significant deviations from the training distribution.
   • CMonitor the model's confidence scores and alert when they drop below a threshold.
   • DPeriodically retrain the model with new data to keep it updated.
   Reveal Solution Discussion Select an answer first
4. 8expert · hard

   A company's AI model that predicts customer churn is found to have a backdoor that triggers when a specific customer ID is present, causing the model to output a high churn probability. The backdoor was likely inserted during training. The company must respond without disrupting ongoing marketing campaigns. Which response is most appropriate?

   • AImmediately shut down the model and switch to manual predictions for all customers.
   • BDeploy a rule-based override that ignores the model's output for the specific customer ID.
   • CIncrease monitoring to detect when the backdoor is triggered.
   • DQuarantine the model, analyze the training data to identify the poisoned samples, and retrain the model after removing them.
   Reveal Solution Discussion Select an answer first
5. 9foundation · easy

   Which of the following is an AI-specific asset that should be included in a threat model?

   • AThe company's marketing brochures
   • BThe employee's personal smartphones
   • CThe trained model weights
   • DThe office building's physical security cameras
   Reveal Solution Discussion Select an answer first
6. 10application · medium

   A retail company's AI-based recommendation system is being assessed for vulnerabilities. The security team has limited time and must prioritize the assessment. Which approach is most effective for identifying high-risk vulnerabilities in the AI system?

   • AReview the model's training data for signs of poisoning.
   • BPerform a threat modeling exercise to identify attack surfaces and then conduct targeted penetration testing on those surfaces.
   • CRun a standard vulnerability scanner on the web application that hosts the recommendation system.
   • DAsk the model developers to list known issues and fix them.
   Reveal Solution Discussion Select an answer first