Utilization of AI-Driven Smart Prompts in Academic Research in Higher Education Institutions

This study investigates how preservice science teachers utilize AI-driven smart prompts to support academic research writing within a digitally mediated learning environment. It explored not only the AI tools used but also the specific aspects of prompt use, including their structure and their functions. Grounded in the Technological Pedagogical Content Knowledge (TPACK) framework, the research employs a qualitative descriptive design to investigate the interactive processes and pedagogical implications of integrating AI. Data were gathered from 65 preservice science teachers (BSED Science) through AI interaction logs, guided reflections, interviews, and research output assessments. Results highlight that AI tools, most notably ChatGPT, Perplexity AI, and Gemini, were extensively used for content generation, methodological guidance, data interpretation, and refining academic language. The study identified diverse prompting strategies (e.g., contextual, iterative, role-based) and mapped seven contextual factors influencing AI use. While AI served as a cognitive and metacognitive scaffold, concerns emerged regarding over-reliance, academic integrity, and the need for critical oversight. Findings underscore the importance of embedding AI literacy and smart prompt engineering into preservice teacher education to promote responsible, student-centered engagement with emerging technologies. The study contributes to the discourse on how AI-mediated learning can transform instructional design and assessment in science education.