Network modeling of science teachers’ perceptions of artificial intelligence integration in science curriculum development

Abstract This study employs psychological network modeling to examine the interrelationships among science teachers’ perceptions of benefits, challenges, and attitudes toward integrating artificial intelligence (AI) tools in science curriculum development. A total of 691 Saudi science teachers completed a validated questionnaire. Network estimation using regularization and centrality analyses identified key themes in science teachers’ perceptions of benefits, challenges, and attitudes toward integrating artificial intelligence (AI) tools into curriculum development and instructional practices. The estimated network revealed a dense structure. Centrality analyses indicated that curriculum content generation and innovative pedagogical tools were most central, while data processing challenges and algorithmic bias emerged as prominent structural barriers. Centrality metrics identify AI-facilitated curriculum design and innovative pedagogical tools as primary benefits, while challenges such as data processing limitations and algorithmic biases emerge as significant barriers. Teacher attitudes demonstrate ambivalence, characterized by resistance to automating administrative functions alongside guarded optimism for interactive and adaptive learning applications. These insights underscore the imperative to address technical constraints, ethical considerations, and the evolving professional identities of educators to ensure effective AI integration. The study contributes a nuanced, systemic understanding that informs the development of tailored AI tools and support frameworks, ultimately advancing transformative innovation in science education.