Identify Guiding Principles for Responsible AI AI-900 Practice Questions (Page 1)

Questions 1–5

1. 1expert · hard

   A government agency uses an AI system to determine eligibility for social benefits. The system sometimes makes errors that could deny benefits to eligible individuals. The agency wants to ensure accountability while also maintaining efficiency. What is the best approach?

   • ALimit the AI system's use to low-stakes decisions and use humans for high-stakes ones.
   • BUse the AI system only as a recommendation and have a human make all final decisions.
   • CPublish the AI system's decision criteria to the public to ensure transparency.
   • DEstablish an appeals process where affected individuals can request a human review of the AI's decision.
   Reveal Solution Discussion Select an answer first
2. 2application · medium · select all that apply

   A company is developing an AI-powered virtual assistant for its website. The company wants to ensure the assistant is inclusive and accessible. Which of the following actions should the company take? Select all that apply.

   • AEnsure the assistant is compatible with screen readers and keyboard navigation.
   • BTest the assistant with users who have various disabilities.
   • CUse only technical jargon to sound professional.
   • DProvide text-based and voice-based interaction options.
   Check answers Discussion Select an answer first
4. 3application · medium

   A bank uses an AI system to approve personal loans. During testing, the model approves a higher percentage of applicants from one demographic group than from another, even when income and credit history are similar. The bank must address this before deployment. What should the bank do first?

   • AEvaluate the model's performance across demographic groups using fairness metrics and inspect the training data for bias.
   • BRemove demographic attributes from the dataset and retrain the model.
   • CDeploy the model and monitor its outcomes over time to see if the disparity persists.
   • DRetrain the model with more data from the group that has the lower approval rate.
   Reveal Solution Discussion Select an answer first
5. 4expert · hard

   A city deploys an AI system to allocate emergency response resources. The system sometimes makes errors that could delay responses. The city must ensure accountability for the system's actions. What should the city do?

   • ALet the system operate autonomously and review its performance annually.
   • BPublish the system's error logs publicly to ensure transparency.
   • CDefine clear roles and responsibilities for monitoring the system and responding to its errors.
   • DDisable the system and use manual dispatch to avoid errors.
   Reveal Solution Discussion Select an answer first
6. 5application · medium · select all that apply

   A company uses an AI system to process customer support tickets. The system is hosted in the cloud and contains sensitive customer information. The company wants to secure the AI system. Which of the following actions should the company take? Select all that apply.

   • ARegularly update and patch the AI system's underlying software.
   • BUse multi-factor authentication for all users who access the system.
   • CStore all customer data in a publicly accessible database for easy access.
   • DMonitor logs for suspicious activity and set up alerts.
   Check answers Discussion Select an answer first