Digital Word-of-Mouth and Teaching Norms in the Adoption of AI Chatbots among University Students in Northern Peru

Abstract

The main objective of this study was to examine how electronic word-of-mouth (WOM) and teacher-related social norms (SN) influenced behavioral intention (BI) and actual use behavior (USEB) of generative AI chatbots among university students in northern Peru, while testing BI as a mediator and sex as a moderator. A quantitative, non-experimental, cross-sectional, explanatory design was employed. Data were collected through a structured Likert-scale questionnaire (1–5) from 520 students in Trujillo, Piura, and Lambayeque, and the proposed model was estimated using PLS-SEM with 5,000-bootstrap resampling. The findings indicated moderately high levels of WOM, SN, BI, and USEB, alongside adequate reliability and convergent/discriminant validity. WOM positively and strongly predicted BI (β=0.56;p<0.001), whereas SN also increased BI but with a smaller effect (β=0.24; p<0.001). BI, in turn, showed a robust association with USEB (β=0.71; p<0.001). Significant indirect effects supported mediation for WOM→USEB (β=0.40) and SN→USEB (β=0.17). Sex significantly moderated the WOM→BI path but did not yield a conclusive moderation for SN→BI. Overall, adoption was concluded to be a socially mediated process, primarily driven by peer communication.