As autonomous vehicles (AVs) become more prevalent, the absence of traditional driver-to-human communication (e.g., gestures and eye contact) may hinder the ability of other road users to infer vehicle intentions, potentially disrupting traffic flow. Consequently, external human–machine interfaces (eHMI) have been proposed for conveying AV intentions. Although previous studies have focused on pedestrian interactions, few have examined the impact of eHMI on the surrounding drivers. This study investigated the effects of eHMI content on driver behavior and situation awareness (SA) in a virtual environment simulating urban traffic using autonomous buses. A total of 20 licensed drivers participated in driving tasks under two scenarios: a pedestrian crossing in front of a stopped bus, and a bus waiting to turn right while another vehicle proceeded straight. The participants experienced both eHMI-present and eHMI-absent conditions. Results demonstrated that eHMI significantly improved SA and gaze behavior, particularly when displaying context-specific messages (e.g., ‘Pedestrian Crossing’ and ‘Oncoming Vehicle’). However, generic messages (e.g., ‘Pay Attention to Traffic’) did not enhance SA and were often disregarded. Although the informative eHMI improved awareness, it also focused on the bus, potentially reducing attention to other elements. These findings highlight the importance of designing informative and contextually relevant eHMI content.