Iranian In-Service Teachers’ Perceptions of Using Generative Artificial Intelligence (ChatGPT) in Instructional Design and Assessment

This study investigates Iranian in-service educators' perceptions of integrating generative artificial intelligence, specifically ChatGPT, into instructional design and assessment practices within Iran's centralized, examination-oriented educational system. Employing a descriptive survey design guided by the Technology Acceptance Model (TAM) and Diffusion of Innovations Theory, data were collected online from ۲۴۸ in-service teachers (۵۷.۳% female, mean teaching experience ۱۲.۴ years, SD = ۵.۶) using a structured questionnaire analyzed through descriptive statistics, inferential tests (ANOVA, t-tests, regression), and thematic coding of open-ended responses. Results revealed moderately positive overall perceptions (M = ۳.۸۲, SD = ۰.۷۸ on a ۵-point Likert scale), with stronger agreement for instructional design (M = ۴.۰۱, ۷۲.۶% agree) than assessment (M = ۳.۶۵, ۵۹.۳% agree). Teachers highly valued ChatGPT for generating feedback (۷۸.۲% agreement) and automating objective grading (۵۹.۳%), but expressed skepticism toward plagiarism detection (۴۲.۷%). Facilitators included potential time savings (۵۲.۰%) and AI training hours (β = ۰.۳۲, p < ۰.۰۰۱), while barriers comprised lack of reliable internet access (۴۵.۲%) and cultural concerns (β = -۰.۱۹, p = ۰.۰۰۴), explaining ۱۸.۴% of variance in perceptions. Demographic influences showed experienced teachers and females holding more favorable.