본문 바로가기
  • Home

How ChatGPT Self-Efficacy and User Characteristics Influence Information-Seeking Behavior via Performance and Effort Expectations: Moderating Role of Interface Concerns

  • Industry Promotion Research
  • Abbr : IPR
  • 2025, 10(4), pp.73~88
  • DOI : 10.21186/IPR.2025.10.4.073
  • Publisher : Industrial Promotion Institute
  • Research Area : Interdisciplinary Studies > Interdisciplinary Research
  • Received : September 16, 2025
  • Accepted : October 16, 2025
  • Published : October 31, 2025

Hyung-June Kim 1 Young Wook Seo 1

1대전대학교

Accredited

ABSTRACT

This study examined the effects of ChatGPT self-efficacy and characteristics on information-seeking behavior through performance expectancy and effort expectancy, and verified the moderating effect of ChatGPT interface concerns. The research model was validated using data collected through a survey of 300 office workers across various industries in Korea from August 5 to 14, 2025, with analysis conducted using SmartPLS. The main findings are as follows. First, ChatGPT self-efficacy had a significant effect on both effort expectancy and performance expectancy, with a particularly stronger influence on effort expectancy. Second, ChatGPT characteristics significantly influenced both performance expectancy and effort expectancy, with a greater effect on performance expectancy. Third, both performance expectancy and effort expectancy had significant effects on information-seeking behavior, confirming the mediating role of the technology acceptance model. Fourth, interface intrusion concerns significantly moderated the relationship between effort expectancy and information-seeking behavior, but the moderating effect with performance expectancy was not significant, confirming only a partial moderating effect. These findings suggest that enhancing users' self-efficacy and understanding of system characteristics is important when implementing generative AI, and that strategies to mitigate interface intrusion concerns are essential for effective utilization of ChatGPT.

Citation status

* References for papers published after 2024 are currently being built.