ABSTRACT This study examines how socially compliant robot behaviours, operationalized as body language and verbal cues, affect human perceptions of social intelligence and compliance with a quadruped robot during incidental human–robot encounters. Extending prior work that primarily focused on direct human–robot interactions, this study addresses the understudied context of incidental encounters, which are brief, unplanned interactions between bystanders and autonomous robots in public environments. In an online video‐based study with 385 participants using a within‐subject design, we found that both verbal and body language behaviours significantly improved the robot's perceived social intelligence (PSI), which in turn positively correlated with human compliance. Verbal communication alone increased compliance likelihood more than body language, and combining both yielded the highest PSI ratings and compliance scores. To validate these findings beyond controlled video scenarios, a follow‐up real‐world study with 26 participants replicated the results: 11 out of 13 of participants complied in the Body Language + Verbal condition versus only 1 in the baseline condition. Free‐text responses revealed the importance of clearly stated intentions, politeness, and concerns about robot legitimacy and safety. Together, these findings highlight the critical role of perceived social intelligence in fostering human assistance for robots during incidental encounters and offer design strategies to support robot deployment in public environments.

Following the development of artificial intelligence, robots are becoming more and more autonomous and increasingly used in public settings, and naturally, with the phenomenon comes new types of human-robot interactions (HRI), such as incidental Human-Robot encounters (HRE). As an emerging research area of HRI, HRE extends and broadens the work of HRI by exploring how users and non-users react to a robot that they incidentally encounter. The present paper reports the results of a work-in-progress project that aims to enrich this field by conducting a field observation of how pedestrians or passersby react to a quadruped robot when the robot is seeking assistance from them in order to enter a building. The findings from the observation provide insights into how robots can be designed to integrate with the human world more smoothly.

ABSTRACT The emergence of AI‐driven technologies often necessitates the collection of private user information to deliver personalised services and enhance the overall user experience. Given the recurring incidents of data breaches, awareness of privacy risks and concerns about disclosing personal information to AI‐driven applications has significantly increased. Privacy concerns have become a critical issue, heavily influencing users' intentions to interact with such systems. To appropriately investigate the antecedents and consequences of disclosing private information, this study examines the influence of social presence (humanlike vs. non‐humanlike media) on privacy concerns and information disclosure across different types of data sensitivities (including retail, financial, and medical data). An online survey (N = 282) and a lab experiment (N = 70) were conducted, incorporating multiple experimental tasks under various conditions. The results reveal that both social presence and data sensitivity significantly impact privacy concerns and information disclosure. Additionally, a privacy paradox is observed: while participants express concern about privacy, their attitudinal and behavioural intentions shift, indicating a willingness to trade sensitive information for enhanced services. The findings also show that individual personality traits strongly influence one's intention to disclose personal information when interacting with humanlike media. Furthermore, when investigating privacy concerns, it is essential to move beyond task‐driven assessments. Instead, identifying the specific types of private information involved and adopting a data‐driven perspective provides a more accurate understanding of privacy‐related behaviours.

The use of humanoid robots has surged in recent decades. However, the nonverbal features in shaping robot personalities remain underexplored. This study investigates how nonverbal cues (including textual and gestural elements) can generate a spectrum of robot personality traits (introvert, ambivert, and extrovert) and evaluates their impact on users’ cognitive perceptions. Textual manipulations involved three iterations, adjusting word count, information structure, and visual effects. Gestural designs underwent two iterations, altering movement frequency, speed, and size. Multiple empirical studies were conducted to assess the development of robot personality traits and their effects. The results confirm the effectiveness of these nonverbal approaches in characterizing diverse robot personalities and significantly influencing users’ cognitive framing. This research provides valuable design guidelines for leveraging a humanoid robot’s nonverbal features to create a variety of personality traits. Our findings emphasize the importance of considering a spectrum of robot personalities rather than focusing solely on extreme traits.

HRI research using autonomous robots in real-world settings can produce results with the highest ecological validity of any study modality, but many difficulties limit such studies’ feasibility and effectiveness. We propose Vid2Real HRI, a research framework to maximize real-world insights offered by video-based studies. The Vid2Real HRI framework was used to design an online study using first-person videos of robots as real-world encounter surrogates. The online study (n=385) distinguished the within-subjects effects of four robot behavioral conditions on perceived social intelligence and human willingness to help the robot enter an exterior door. A real-world, between-subjects replication (n=26) using two conditions confirmed the validity of the online study’s findings and the sufficiency of the participant recruitment target (n=22) based on a power analysis of online study results. The Vid2Real HRI framework offers HRI researchers a principled way to take advantage of the efficiency of video-based study modalities while generating directly transferable knowledge of real-world HRI. Code and data from the study are provided at vid2real.github.io/vid2realHRI.