AI Transparency and Trust Formation: Dual Attribution Path-ways of Ethical Authenticity and Perceived Manipulation

This study examines how transparency functions as a design feature in AI-mediated communication interfaces and how it shapes user trust in intelligent systems. As AI increasingly participates in human-computer interaction, transparency is often implemented to improve system interpretability and trustworthiness. However, user responses to transparency remain psychologically complex. Using an experimental design, this study investigates how transparency influences trust through competing attributional interpretations of system behavior. The results show that transparency can simultaneously enhance perceptions of ethical authenticity while also triggering perceptions of manipulative intent. These opposing interpretations exert differential effects on trust, indicating that transparency does not operate as a uniformly beneficial interface feature.The findings suggest that transparency functions not only as informational disclosure but also as a communicative signal that shapes how users interpret system intentions, responsibility, and influence. From an applied perspective, the results highlight the importance of carefully designing transparency features in AI interfaces to promote perceived legitimacy while minimizing unintended suspicion. This study contributes to research on human-AI interaction by clarifying the psychological mechanisms through which transparency influences trust and by providing design implications for trustworthy intelligent systems.