Functional Stratification of AI Assistance Models in Academic Writing: Student Perceptions in Indonesian Universities

Abstract

AI’s growth has changed many areas, including higher education and the way students write academically. Students’ views of AI help models, on the other hand, are complicated and varied, depending on how they are used. This study focused on how students view AI-assisted models in academic writing. There were 157 students in this exploratory quantitative study. Researchers used Principal Component Analysis to reduce the number of factors in the data and the Rasch model to see how much students agreed on the different sorts of AI assistance. The results of the quantitative analysis indicate that all factor loadings exceed 0.5, suggesting a strong relationship between the variables. The model demonstrates an acceptable to satisfactory fit, though some indices are marginal (χ 2 /df = 2.631; RMSEA = 0.090 [CI: 0.074–0.11]; GFI = 0.860; AGFI = 0.820; CFI = 0.970; NFI/IFI/NNFI = 0.960). Additionally, student perceptions are reported to fall within the logit range of –1.14 to +1.63, reflecting varying degrees of agreement regarding the role of AI in academic writing. The results show that there are two AI function models in academic writing: AI-assisted idea generation and AI-enhanced academic writing quality, which possess 15 components. AI helps students generate ideas, develop content, organize their writing, refine their thoughts, and conduct research, depending on the sources. AI also helps with language accuracy, discourse coherence, and cohesion in texts, and style clarity and enhancements. The findings of this study indicate that students should employ AI not only for the technical components of writing but also to augment introspective, analytical, and argumentative methodologies.

Ähnliche Arbeiten