6.4 Explain Human-In-The-Loop (HITL) Strategies AI-TECHNICAL-PRACTITIONER Practice Questions (Page 1)

Questions 1–5

1. 1foundation · easy

   How does human-in-the-loop (HITL) contribute to the safety of AI systems?

   • ABy ensuring the AI system operates without any human intervention
   • BBy making the AI system faster so it can respond to threats more quickly
   • CBy allowing humans to intervene in risky or uncertain situations
   • DBy reducing the amount of data the AI system needs to process
   Reveal Solution Discussion Select an answer first
2. 2foundation · easy

   Which human role in an AI workflow is responsible for making the final call on a high-stakes action suggested by the AI?

   • AAnnotator
   • BDecision-maker
   • CData engineer
   • DReviewer
   Reveal Solution Discussion Select an answer first
4. 3application · medium

   A healthcare organization is deploying an agentic AI system that drafts patient discharge summaries. The system also suggests follow-up appointments. The organization wants a human to verify the accuracy of the summaries before they are sent to patients. Which human role is most appropriate for this task?

   • ATrainer, to provide feedback to the AI model after each summary.
   • BDecision-maker, to decide which patients receive the summaries.
   • CReviewer, to check the AI-generated summaries for errors and approve them for release.
   • DAnnotator, to label the AI's outputs for future training.
   Reveal Solution Discussion Select an answer first
5. 4application · medium

   A financial services firm deploys an agentic AI system that autonomously executes routine account transactions. To ensure responsible operation, the firm wants to incorporate human oversight without slowing down high-volume, low-risk transactions. Which HITL strategy best meets this requirement?

   • ADeploy the agent without any human checks, relying on post-hoc audits to catch errors.
   • BUse a risk-based escalation model where only transactions above a certain threshold or flagged as anomalous require human approval.
   • CHave a human manually execute all transactions and use the AI only for reporting.
   • DRequire human approval for every transaction the agent initiates.
   Reveal Solution Discussion Select an answer first
6. 5application · medium

   A retail company uses an agentic AI to manage product pricing. The AI adjusts prices based on demand and inventory. The company wants to ensure that price changes are not too aggressive, which could upset customers. Which HITL strategy is most suitable?

   • ALet the AI adjust prices freely without human review.
   • BHave a human review and approve all price changes before they are applied.
   • CSet a threshold for price changes; changes beyond the threshold require human approval, while smaller changes are applied automatically.
   • DHave a human manually set all prices based on the AI's recommendations.
   Reveal Solution Discussion Select an answer first