Question 36
Domain 4: Guidelines for Responsible AIA product team is redesigning an AI assistant to make its outputs easier for end users to understand and trust. Which approach best reflects human-centered design for explainable AI?
Correct answer: B
Explanation
Human-centered explainable AI should be designed to help people understand AI decisions and provide feedback, because transparency and feedback loops make systems more usable and trustworthy. — 50_20_Governance_and_compliance_for_AI.md
Why each option is right or wrong
A. Limit user interaction with the system so outputs are accepted consistently without additional explanation
Human-centered design emphasizes user-feedback mechanisms and decision transparency, not reducing opportunities for user input.
B. Provide clear information about how decisions are made and include a way for users to submit feedback on outputs
The source material for Task 4.2 identifies human-centered design for explainable AI as including both user-feedback mechanisms and AI decision transparency. In this redesign scenario, giving users visibility into how the assistant reached outputs and a channel to respond directly matches those two named principles.
C. Focus only on improving model performance metrics because accuracy alone creates user trust
Responsible AI in this domain distinguishes explainability from performance; trust also depends on transparency and feedback loops.
D. Hide decision details behind internal documentation so the system remains simple for users
Transparency in AI decision-making should be surfaced to users, not confined only to internal records.