Assessing a ChatGPT-Integrated EFL Program Through Self-Determination Theory and Technology Acceptance Model

The rapid development of generative AI (GenAI) has significantly impacted language learning, offering personalized support while raising concerns regarding accuracy, ethics, and overreliance. This study investigates how students' basic psychological needs (autonomy, competence, and relatedness) and Technology Acceptance Model (TAM) constructs change following a ChatGPT-supported English as a Foreign Language (EFL) learning program. Data from 34 Japanese university students were analyzed using t-tests, linear mixed-effects (LME) model, and multiple regression analysis. The results revealed significant increases in perceived autonomy, perceived competence, perceived usefulness (PU), attitude (AT), and actual use (AU), while perceived relatedness, perceived ease of use (PE), and behavioral intention (BI) remained unchanged. Initially, GenAI usage frequency influenced PE, BI, and AU; however, perceived competence emerged as a stronger predictor after the program. These findings highlight the need for instructional strategies that enhance students' competence while ensuring ethical and balanced integration of GenAI use in EFL learning.