Exploring primary preservice teachers’ artificial intelligence facilitated metacognition in a natural sciences curriculum course

The rise in the use of artificial intelligence (AI) in higher education by students and instructors raises concerns about how learning happens in these new environments. Most technological advancements come as a disruption to the traditional ways of teaching and learning, and therefore, classroom practices need to be adjusted accordingly. Luckin et al. (2016) concede that AI has transformed education. Given that AI tools can perform tasks traditionally performed by humans, including learning, instructors need to ensure that learning remains human-centered for the development of the targeted knowledge domains and skills in the classrooms. Albus and Seufert (2025) argue that AI must be used to facilitate learning but not replace students’ learning. Similarly, Fan et al. (2025) observed that students can become overly reliant on AI tools in ways that reduce meaningful learning. Despite the identified risks to meaningful learning when students use AI tools, there is evidence that when used appropriately as tools to support learning, this educational technology presents many opportunities. Yu et al. (2024) showed that using educational chatbots in a biology education classroom provided students with metacognitive feedback, enabling meaningful learning. In addition, Atchley et al. (2024) assert that using AI tools in the classrooms supports personalized and self-regulated learning. Personalized learning and self-regulated learning align with learner-centered classroom practices that promote metacognition. In this study, we used Flavell’s model on metacognition to explore the use of AI tools to facilitate meaningful learning for primary preservice teachers attending a natural sciences curriculum course at one South African university as a theoretical framework. Accordingly, the metacognition components of self-regulation and self-reflection were integrated with the affordances of AI to support learning and form a conceptual framework for the study. The AI affordances to support meaningful learning in this study include personalized learning, self-regulated learning, enhanced creativity and problem-solving, collaborative learning, ethical and critical thinking, resource efficiency and generation, and scaffolded inquiry. The study employed a realist research paradigm, a mixed methods approach, and a parallel convergent design to collect data utilizing a survey, document analysis, a reflection schedule, and focus group interviews. For the survey, a close-ended questionnaire based on a five-point Likert scale (strongly agree-5, agree-4, not sure-3, disagree-2, and strongly disagree-1) was applied to 300 primary preservice teachers enrolled in a natural sciences curriculum course and conveniently sampled. The quantitative data were analyzed through descriptive statistics and regression analysis to determine the relationship between the independent variables (collaborative learning, personalized learning, self-regulated learning, and scaffolded inquiry) and the dependent variables (creativity and problem-solving, ethical and critical thinking, self-reflection, and resource efficiency and generation). The qualitative data were collected from 30 purposively selected preservice teachers who were most competent in using AI tools using five focus group interviews, document analysis of a project completed using AI tools, and an open-ended questionnaire to capture the students’ reflections. Directed qualitative content analysis techniques were used to analyze the data. The study's primary research question is, how does AI-facilitated instruction influence the primary preservice teachers’ metacognition in a natural sciences curriculum course? Accordingly, three subsidiary questions were formulated: (1) What AI tools are used to facilitate learning by primary preservice teachers in a natural sciences course? (2) what factors influence the metacognition of primary preservice teachers during AI-facilitated learning in a natural sciences curriculum course? (3) how can metacognition during AI-facilitated learning by primary preservice teachers in a natural sciences curriculum course be explained? The findings of the study showcase the tools used by the primary preservice to achieve metacognition and the factors that influence metacognition during AI-facilitated learning. In addition, the study proffers context-based explanations of how preservice teachers can achieve human-centered learning when AI-based instructional strategies are used. References Albus, P., Seufert, T. (2025). Artificial Intelligence in Education: the Importance of Metacognitive Monitoring. In: Montag, C., Ali, R. (eds) The Impact of Artificial Intelligence on Societies. Studies in Neuroscience, Psychology and Behavioral Economics. Springer, Cham. https://doi.org/10.1007/978-3-031-70355-3_8 Atchley, P., Pannell, H., Wofford, K. et al. Human and AI collaboration in the higher education environment: opportunities and concerns. Cogn. Research 9, 20 (2024). https://doi.org/10.1186/s41235-024-00547-9 Fan, Y., Tang, L., Le, H., Shen, K., Tan, S., Zhao, Y., Shen, Y., Li, X., & Gaševic, D. (2024). Beware of Metacognitive Laziness: Effects of Generative Artificial Intelligence on Learning Motivation, Processes, and Performance. ArXiv. https://doi.org/10.1111/bjet.13544 Flavell, J. H. (1976). Metacognitive aspects of problem solving. In L. B. Resnick (Ed.), The nature of intelligence (pp. 231–236). Lawrence Erlbaum Associates. Flavell, J. H. (1979). Metacognition and cognitive monitoring: A new area of cognitive-developmental inquiry. American Psychologist, 34(10), 906–911. Flavell, J. H. (1992). Cognitive development: Past, present, and future. Developmental Psychology, 28(6), 998–1005. Luckin, R., Holmes, W., Griffiths, M., & Forcier, L. B. (2016). Intelligence Unleashed: An Argument for AI in Education. Pearson Education. Yin J, Zhu Y, Goh T-T, Wu W, Hu Y. Using Educational Chatbots with Metacognitive Feedback to Improve Science Learning. Applied Sciences. 2024; 14(20):9345. https://doi.org/10.3390/app14209345