Case Study / Human-AI Interaction

AI Delegation & Arbitration

A research-led interaction framework for deciding when AI systems should act, defer, negotiate, or escalate decisions to people.

Delegation / arbitration / trust calibration

Overview

This case study examines how teams can collaborate with AI agents without losing agency, accountability, or situational awareness. The work focuses on the seams where autonomous systems need human judgment.

Approach

Layomi mapped decision thresholds, conflict states, and escalation paths, then prototyped interaction patterns that make machine confidence, disagreement, and handoff moments visible.

Outcome

A set of interface principles for AI delegation that treats arbitration as a design material, not just an error state.

Back to selected work