AI literacy and trust: A multi-method study of Human-GAI team collaboration

Abstract

As artificial intelligence (AI) becomes increasingly integrated into team settings for collaboration with humans, understanding the dynamics of trust and AI literacy is essential for enhancing team effectiveness. This study investigates the relationship between trust and AI literacy in human-generative AI (GAI) team collaboration, focusing on how AI literacy affects trust formation in these interactions. Drawing upon foundational teamwork literature and AI literacy frameworks, we conducted a multi-method investigation involving 116 undergraduate team members across 23 project teams throughout a semester. In Study 1, qualitative findings revealed distinct attitudes toward GAI as a teammate, categorized as trust, distrust, and ambivalence. Study 2 employed quantitative methods to determine predictors of trust in GAI, demonstrating that AI knowledge and perceived value—key components of AI literacy—significantly influenced perceptions of trust. Notably, perceptions of GAI accuracy emerged as a critical determinant of trust. Our findings highlight the complex interplay between AI literacy and trust in human-GAI collaboration. We observed a paradox: increased AI literacy can enhance collaboration but may also lead to hesitancy in future AI use. We contribute to advancing the understanding of human-AI collaboration by highlighting the critical role of AI literacy in shaping trust and socio-technical team dynamics. Our study provides evidence demonstrating the importance of targeted AI literacy development in building trust and fostering effective collaboration in human-GAI teams. These findings provide a foundation for research aimed at optimizing human-GAI teamwork and developing adaptive AI literacy frameworks, empowering individuals to effectively engage with AI across diverse collaborative settings.