Executives require that a customer-facing multi-agent assistant expose uncertainty, allow users to correct the agent mid-flow, and avoid deceptive anthropomorphism, while logging consent for data use. Which architectural focus addresses these requirements?
Select an answer to reveal the explanation.
Short Explanation
You need human-AI experience and responsible AI controls: transparent agent identity, uncertainty communication, user correction/override affordances, and consent-aware data handling. That covers the executives' requirements without pretending to be human, hiding confidence, or keeping personal data without consent. Bake these into UX and agent behavioral guidelines at design time, not as a marketing afterthought.
Full Explanation
Architecture must support human-AI experience (HAX) and responsible AI. The correct focus—transparent agent identity, uncertainty communication, user correction/override, and consent-aware data handling—covers transparency, uncertainty, correction, and consent. Claiming the agent is a licensed human expert is deceptive anthropomorphism. Hiding confidence and blocking goal edits reduces user agency and obscures model limits. Indefinite personal-data retention without consent violates privacy and consent expectations. Specify UI patterns, agent system guidelines, and data retention aligned to policy, and evaluate with user studies and compliance review.