Veröffentlichungen von Johannes Bernard

AI increasingly supports interpersonal communication by helping individuals create or refine messages, raising the question of whether such usage should be disclosed to message receivers. While prior research on AI usage disclosure has focused on organizational and academic contexts shaped by regulations, disclosure in interpersonal communication remains voluntary and underexplored. This research-in-progress paper investigates which relational patterns shape AI usage disclosure in AI-supported interpersonal communication. Drawing on an exploratory qualitative study based on five semi-structured interviews, we aim to identify relational patterns structured by communication type and receiver type. Our preliminary results suggest that AI usage disclosure is perceived as irrelevant in functional communication with distant receivers, permissible in functional communication with close receivers, precarious in intimate communication with distant receivers, and improbable in intimate communication with close receivers. We will contribute to literature by conceptualizing AI usage disclosure as a relational judgment.

Recommender systems on social media increasingly personalize information, creating information-limiting environments that may shape how users evaluate social media platforms and whether they intend to switch. This research-in-progress paper examines how two characteristics of such environments, information slant and information diversity, influence user satisfaction and switching intentions on social media. Drawing on switching literature, we argue that both opinion-congruent information and diverse information can – whilst being opposingly associated with information-limiting environments – increase satisfaction, which in turn reduces switching intentions. We further propose that truth bias, critical consuming media literacy, and issue involvement moderate these relationships. We will analyze data from Twitter (now X) users using partial least squares structural equation modeling. Exploratory results from a pre-test indicate that perceived information slant and perceived information diversity positively influence satisfaction, while satisfaction negatively affects switching intentions. We will contribute to literature by linking information-limiting environments and their characteristics to social media switching.

Physiological videoconference fatigue (PVCF) has emerged as a significant threat to employee well-being. While current research focusses on the drivers of PVCF, this study integrates expectancy-violations theory to examine how prior expectations regarding videoconferences influence subsequent physiological recovery to homeostasis. We conducted a 2x3 between-subject laboratory experiment and measured PVCF through skin conductance within a simulated videoconference environment. Preliminary results indicate that higher amplitudes in physiological arousals do not necessarily lead to longer recovery times and that violated expectations lead to longer recovery times than confirmed expectations.