From Assistance to Affective Co-Driving: Novice Drivers’ Emotional Needs and Design Expectations for In-Vehicle AI Agents
TL;DR
HACoD is proposed, a Human–Agent Emotional Co-Driving Framework that links driver emotional needs, human–agent collaboration patterns, and affective agent design elements and highlights the need for transparent data boundaries and real-world validation.
Abstract
As AI becomes increasingly embedded in intelligent cockpits, human-vehicle interaction is shifting from task assistance toward affective and collaborative driving experiences. Yet little is known about what emotional support drivers expect from in-vehicle AI agents or how these expectations can inform design. We conducted semi-structured interviews with 22 novice drivers and a follow-up evaluation with four experts from relevant domains. Thematic analysis identified six design dimensions spanning interaction modality, agent persona, feedback strategy, contextual adaptation, system transparency, and personalization. Participants expected AI agents to combine safety-critical guidance with emotionally calibrated support, including adaptive voice tone, multimodal prompts, explainable interventions, privacy-aware sensing, and roles that evolve from coach to companion. Based on these findings, we propose HACoD, a Human–Agent Emotional Co-Driving Framework that links driver emotional needs, human–agent collaboration patterns, and affective agent design elements. Expert feedback supported the framework’s coherence and design applicability while highlighting the need for transparent data boundaries and real-world validation.
