ScholarWorks@중앙대학교

초록

Generative AI technologies based on large language models (LLMs) offer significant advantages in natural language processing and context-aware response generation, enabling seamless and human-like interactions with users. This study investigates how the perceived anthropomorphism of ChatGPT affects user trust and privacy concern. Grounded in privacy calculus theory and the uses and gratifications theory, the research further examines whether user interaction gratification moderates privacy concern. A survey was conducted among 152 university students who used ChatGPT for various academic purposes. Using PLS-SEM for hypothesis testing, the results reveal that anthropomorphic perception of ChatGPT significantly enhances trust while also influencing privacy concern. Additionally, user interaction gratification(content) with ChatGPT was found to alleviate privacy concern. This study contributes to a deeper understanding of the psychological mechanisms in human-AI interactions and offers empirical insights for designing user-centered AI services in the era of generative AI.

키워드