Beyond Technology Acceptance: An Interplay of Self-Efficacy, Language Proficiency, and ChatGPT Adoption from a TAM Perspective

Fauzi, Chandra
Rahmani, Eka Fajar
Utimadini, Nindya Juwita

Jurnal Pendidikan Progresif
Vol 15, No 3 (2025): Jurnal Pendidikan Progresif

Beyond Technology Acceptance: An Interplay of Self-Efficacy, Language Proficiency, and ChatGPT Adoption from a TAM Perspective. Objective: Adopting Generative AI tools for writing improvement, such as ChatGPT, among EFL students is a complex phenomenon where perceptions of technology, often explained by the Technology Acceptance Model (TAM), intersect with important psychological factors like self-efficacy as the theoretical lens. This study seeks to empirically examine the constructs of the TAM framework and scrutinize the differences in ChatGPT acceptance viewed from students' self-rated writing proficiency. Method: This quantitative research adopted a descriptive-correlational design, utilizing convenience sampling to recruit 76 fourth-semester English Education EFL learners at a state university in West Kalimantan. A 29-item questionnaire was administered to collect the data, which was subsequently analyzed using the Mann-Whitney U test, Spearman correlation, and multiple regression. Findings: The regression model predicting attitude was significant (F (2,73) = 161.65, p < 0.001), in which 81.1% of the variance (Adjusted R2 = 0.811) was accounted for by PU (Perceived Usefulness) and Perceived Ease of Use (PEOU), confirming PU as the main predictor (β = 0.644). At the same time, PEOU also exerted a notable influence (β = 0.301), and a strong positive correlation was found between attitude and actual usage (r = .845, p < 0.01). A significant difference in attitude existed between writing proficiency groups (p = .039), with the intermediate group showing more positive attitudes. Conclusion: The results demonstrate that whereas TAM constructs strongly predict technology acceptance, they are also influenced by students' writing proficiency. Furthermore, a significant difference exists in ATU, but not in BI, in terms of writing proficiency. These findings highlight the necessity for differentiated teaching practices that recognize discrepancies in students' writing proficiency and the potential of self-efficacy level. Keywords: academic writing, ChatGPT, Self-efficacy, technology acceptance model (TAM), writing proficiency.