An Alternative Method for Assessing Transparency in Semi-autonomous Vehicles: A Directed Content Analysis of User-Generated Content

This study investigates Transparency levels in modern semi-autonomous vehicles by evaluating Human-Machine Interface (HMI) communication through the lens of Situation Awareness-based Agent Transparency (SAT). Through Directed Content Analysis (DCA), we analyzed 93 User-Generated YouTube videos across eight major vehicle manufacturers, employing structural qualitative coding to understand which specific HMI features are used the most to communicate ADAS status and rationale, and which levels of transparency are reached by different car brands across their models. Findings have revealed that 67% of interactions are restricted to SAT Level 1 (status notification), communicating mostly through visual outputs in the graphic user interface. This research underscores the need for standardized HMI architectures to ensure safe human-machine collaboration as the industry pivots toward SAE Level 3 automation.