Privacy, Ethical, Trust and Safety Controls AAISM Practice Questions (Page 5)

Questions 21–25

1. 21foundation · easy

   Which approach to AI system design is most directly associated with providing clear explanations for individual decisions made by the model?

   • APost-hoc interpretability
   • BModel compression
   • CEnsemble methods
   • DData augmentation
   Reveal Solution Discussion Select an answer first
2. 22foundation · easy

   Which human oversight mechanism is used to allow a person to review and override the decisions made by an AI system before they are finalized?

   • AFeature engineering
   • BAutomated scaling
   • CData deduplication
   • DHuman-in-the-loop
   Reveal Solution Discussion Select an answer first
4. 23application · medium

   An autonomous vehicle company is testing an AI system that controls braking in emergency situations. During testing, the system occasionally fails to detect pedestrians in low-light conditions. The company wants to implement safety controls to prevent harm. Which control should be implemented first?

   • AIncrease the sensitivity of the pedestrian detection algorithm to catch more cases.
   • BRequire a human driver to be ready to take over at all times.
   • CImplement a fail-safe mechanism that defaults to braking when the system is uncertain.
   • DLimit the vehicle's speed in low-light conditions to reduce the impact of a collision.
   Reveal Solution Discussion Select an answer first
5. 24expert · hard

   A hospital is using an AI system to recommend medication dosages for patients. The system has been shown to be highly accurate, but there have been a few cases where it recommended a dosage that was too high, causing adverse effects. The hospital wants to maintain the system's benefits while ensuring accountability and safety. What is the most appropriate action?

   • ADiscontinue the use of the AI system and rely solely on human expertise.
   • BImplement a human-in-the-loop process where a pharmacist reviews and approves all AI dosage recommendations before administration.
   • CAllow the AI system to administer dosages automatically, but log all decisions for later review.
   • DAdd a warning label to the AI system's output stating that dosages may be incorrect.
   Reveal Solution Discussion Select an answer first
6. 25foundation · easy

   Which safety control is designed to automatically stop an AI system from operating if it detects that the system is behaving outside its intended parameters?

   • AKill switch
   • BAudit logging
   • CData encryption
   • DAccess control
   Reveal Solution Discussion Select an answer first