Artificial Intelligence in Learning Design: Acceptance, Perceived Effectiveness, and Barriers

Abstract

This study mapped perceptions of AI in learning design in the Educational Technology Study Program at Universitas Negeri Surabaya (UNESA). A 25-item, 5-point Likert questionnaire (acceptance, perceived effectiveness, limitations; TAM-informed) was completed by 16 lecturers and 130 students selected purposively (users of, or strongly interested in, AI). Content validity met conventional thresholds (all I-CVI ? 0.78; S-CVI > 0.90). Agreement on acceptance (10 items) averaged 82.6% for students (range 80.0–88.5%) and 85.0% for lecturers (range 81.25–87.5%). Agreement on perceived effectiveness (8 items) averaged 85.4% for students (range 80.8–89.2%) and 87.5% for lecturers (range 81.25–93.75%), indicating that respondents believe AI can accelerate material preparation, support adaptive/diagnostic feedback, and enable more personalized learning. Limits were also evident (7 items): difficulty understanding AI (65.4% students; 62.5% lecturers), context relevance of AI outputs (58.5%; 62.5%), curricular alignment (56.9%; 56.3%), feeling safe sharing data (53.9%; 56.3%), and LMS integration (60.8%; 68.8%). Reported training was uneven (61.5% students; 68.8% lecturers), implying roughly 32–38% lacked training. Given the single-site, descriptive design, findings are self-reports—not causal or broadly generalizable. Implications point to pilot-first adoption, targeted capacity building, clearer privacy/ethics governance, and infrastructure alignment before any scale-up.