<![CDATA[This study examines and compares the effects of ChatGPT and Scite_AI, on plagiarism tendencies among students of Islamic Religious Education in Indonesia. Adopting a quantitative research design, the study employed multiple linear regression analysis to evaluate both the partial and simultaneous influences of these tools on academic plagiarism. Prior to regression analysis, classical assumption tests—including normality (Prior to the regression analysis, classical assumption tests—including normality (Kolmogorov-Smirnov p = 0.088), multicollinearity (VIF < 10), multicollinearity (VIF < 10), heteroscedasticity (Breusch-Pagan test), and linearity (scatterplot of residuals)—were rigorously conducted to ensure model validity. The results reveal that both AI tools significantly contribute to increased plagiarism tendencies; however, ChatGPT demonstrates a markedly stronger effect (β = 0.4941; p < 0.001) compared to Scite (β = 0.1042; p < 0.001). The overall regression model is statistically significant (F = 87.32, p = 0.000) and satisfies all classical assumptions, confirming its reliability. Theoretically, this research enriches academic integrity literature by positioning AI tool typology—particularly the distinction between generative and verification tools—as a critical predictor of plagiarism behavior. Practically, it calls for differentiated AI literacy strategies in Islamic higher education, advocating for the integration of adab al-‘ilmu (ethics of knowledge) into digital literacy curricula to foster moral discernment and responsible technology use among future religious educators.]]>
Copyrights © 2025
