Yapay Zeka dil öğretmeni olarak: Ürdünlü üniversite öğrencilerinin İngilizce'ye yönelik yapay zeka tarafından üretilen geri bildirim algıları

Özet (abstract)

This study examines Jordanian university students’ evaluations of AI-generated feedback on their English and the extent to which these evaluations predict continued-use intention. A questionnaire was completed by 750 students, with analysis of the 25 perception items restricted to the 720 respondents with prior AI experience. The items assessed feedback quality, personalization, usefulness, trust, motivation, and continued-use intention. Exploratory and confirmatory factor analyses supported a six-factor structure. Students rated usefulness (M = 4.00) and continued-use intention (M = 3.98) more favorably than trust (M = 3.39) and personalization (M = 3.50). Feedback quality correlated positively with trust (r = .35), while personalization correlated with motivation (r = .41). Usefulness, trust, and motivation jointly predicted continued-use intention, explaining 14.8% of its variance. Higher self-assessed English proficiency was associated with more favorable evaluations of feedback quality (η² = .045). Formal AI training was associated with slightly higher perceived usefulness (d = 0.19), and frequency of use showed a weak positive association with continued-use intention (ρ = .14). Perceived accuracy had little association with usefulness or intention, indicating that students’ appreciation of AI feedback did not necessarily reflect confidence in its correctness. These findings support integrating explicit instruction in evaluating AI-generated suggestions into university English programs, retaining teacher guidance and providing additional support for less proficient learners.

Yazarların özeti; kaynağından alınmıştır. Research Journal in Advanced Humanities, 2026 · DOI ↗