The success or failure of human-robot interaction (HRI) is influenced by various human-centered factors, including personality traits. These variables shape users’ perceptions, engagement, and trust in robotic systems. One key concept in this domain is the similarity hypothesis, which posits that people are more likely to respond positively to robots that exhibit traits or behaviors similar to their own. Understanding how personality alignment affects HRI outcomes is crucial for designing effective and socially compatible robotic agents. While the similarity hypothesis has found support in, e.g., assistive healthcare contexts, its applicability to driving scenarios remains underexplored. To address this research gap, we used the CARLA simulator to create a driving scenario with the humanoid iCub robot endowed with either an extraverted or intraverted personality expressed through non-verbal cues (e.g., motion, gaze) and speech characteristics (e.g., pitch, frequency). Although robot personality was successfully manipulated, participants perceived the robot as more extraverted and preferred it as a driving assistant rather than a social companion. Perceived similarity in extraversion influenced their preferences, partially supporting the similarity hypothesis. Regardless of assigned personality, the robot was seen as more machine-like than human-like. The intraverted robot elicited greater cognitive trust, suggesting confidence in its functional reliability. Participants’ ability to imagine the driving scenario with a robot significantly shaped their perceptions of HRI, underscoring the role of repeated interactions in developing trust and social bonds in HRI.