Identify Guiding Principles for Responsible AI AI-900 Practice Questions (Page 4)

Questions 16–20

1. 16expert · hard

   A hospital uses an AI system to recommend medication dosages. The system occasionally makes errors that could harm patients. The hospital wants to reduce the risk of harm while still benefiting from the AI's efficiency. What is the best approach?

   • ADisable the AI system entirely to avoid any risk of harm.
   • BRequire a human clinician to review and approve all AI recommendations before they are implemented.
   • CUse the AI system only for low-risk patients and rely on clinicians for high-risk cases.
   • DIncrease the AI's confidence threshold so it only makes recommendations when it is highly certain.
   Reveal Solution Discussion Select an answer first
2. 17application · medium · select all that apply

   A health app uses AI to provide personalized fitness recommendations. The app collects sensitive health data from users. The company wants to protect user privacy. Which of the following actions should the company take? Select all that apply.

   • AShare health data with third-party advertisers to improve recommendations.
   • BEncrypt health data both at rest and in transit.
   • CObtain explicit consent from users before collecting their health data.
   • DAllow users to access, correct, and delete their personal data.
   Check answers Discussion Select an answer first
4. 18application · medium

   A retail company uses an AI chatbot to handle customer service. The chatbot collects personal data such as names, addresses, and purchase history. The company wants to ensure privacy while still providing personalized recommendations. What should the company do?

   • AAllow the chatbot to share personal data with third-party vendors to enhance recommendations.
   • BCollect as much data as possible to improve the chatbot's accuracy.
   • CAnonymize or pseudonymize personal data before using it to train or personalize the chatbot.
   • DStore all personal data in an encrypted database and never delete it.
   Reveal Solution Discussion Select an answer first
5. 19expert · hard

   A company deploys an AI model that predicts equipment failures. The model is hosted in the cloud and accessed by field technicians via a mobile app. The company is concerned about adversarial attacks that could manipulate the model's inputs to cause incorrect predictions. What is the best defense?

   • AUse a more complex model that is harder to attack.
   • BImplement input validation and anomaly detection to reject suspicious or manipulated inputs.
   • CRegularly retrain the model with new data to improve accuracy.
   • DRestrict access to the model by requiring a VPN for all technicians.
   Reveal Solution Discussion Select an answer first
6. 20application · medium · select all that apply

   A company uses an AI system to recommend marketing strategies. The system sometimes makes recommendations that could be harmful to the company's reputation. The company wants to ensure accountability. Which of the following actions should the company take? Select all that apply.

   • AAssign a specific team or individual to oversee the AI system's recommendations.
   • BDocument all decisions made by the AI system and the rationale behind them.
   • CLet the AI system operate without any human oversight to avoid bias.
   • DEstablish a process for reviewing and overriding AI recommendations when necessary.
   Check answers Discussion Select an answer first