Asking AI About Your Car: AI-Augmented Vehicle Owner’s Manuals Support Integrated System Understanding Beyond Traditional Documentation

ObjectiveThis study investigates the effectiveness of an AI-powered interactive vehicle owner's manual compared to a traditional static manual in improving users' understanding of Advanced Driver Assistance Systems (ADAS) using a production vehicle's owner's manual.BackgroundAs vehicle automation becomes increasingly complex, drivers face challenges in understanding ADAS features. While traditional owner's manuals have demonstrated effectiveness when they are used, there remains potential to enhance driver engagement through AI and provide more interactive and accessible learning experiences.MethodsUsing a between-subjects design, 38 participants were randomly assigned to learn about four commercially available ADAS features using either a PDF manual or a Retrieval-Augmented Generation (RAG) AI manual. Mental model accuracy was assessed through multiple-choice questions, while participants' reasoning patterns were analyzed using structural topic modeling (STM) of open-ended responses.ResultsBoth training methods improved mental model accuracy from pre- to post-training, with no significant differences between PDF and RAG conditions in quantitative learning outcomes. However, STM analysis revealed distinct qualitative differences in the participants' reasoning patterns. RAG-trained participants demonstrated more sophisticated systems-level thinking, particularly in feature integration reasoning. Analysis through the lens of the Technology Acceptance Model revealed that both methods operate through similar psychological mechanisms, with perceived usefulness aiding user acceptance.ConclusionAI-augmented owner's manuals achieve comparable learning effectiveness to traditional documentation while enhancing feature integration reasoning. Interactive AI systems serve as effective enhancements rather than replacements for proven educational approaches, guiding users toward more sophisticated mental models of complex automated systems.ApplicationThis research provides insights for automotive manufacturers and documentation specialists on effective approaches for educating drivers about complex vehicle automation systems, potentially improving safety and user experience.