The Shift from HMI to HAIC

Human-Machine Interaction (HMI) and Human-AI Collaboration (HAIC) require fundamentally different evaluation approaches.

HMI focuses on usability: Can the user accomplish their task? How efficiently? With how few errors? The machine is a tool; the human is the actor. Evaluation measures interface quality and task performance.

HAIC focuses on collaboration quality: How well do human and AI work together? Who contributes what? How do they adapt to each other? Both are actors; the outcome is joint. Evaluation must capture the dynamics of partnership.

The distinction matters because applying HMI metrics to HAIC misses crucial dimensions: trust development, mutual adaptation, complementary contribution, and the emergent capabilities that arise from true collaboration.

Traditional usability frameworks (Nielsen, Norman) assume the human is trying to use a system. HAIC assumes human and AI are trying to solve problems together. The relationship is different. The measurement must be different.

Related: 05-atom—haic-three-modes, 07-molecule—evaluation-methods-tradeoff