Promoting Interdisciplinary Learning With Generative AI Through Self‐Regulated Scaffolding

ABSTRACT Background Interdisciplinary learning in higher education is increasingly influenced by Generative AI (GenAI), which strengthens disciplinary foundations. However, effectively engaging students with GenAI to actively manage their learning processes while maintaining motivation and adaptability remains a significant challenge. Self‐regulated scaffolding is recognised as an effective tool for helping students navigate the three phases of self‐regulated learning (SRL): forethought, performance, and self‐reflection. SRL is closely associated with the fulfilment of three basic psychological needs: autonomy, competence, and relatedness. Objectives This study introduced a self‐regulated scaffolding approach based on self‐determination theory (SDT) and evaluated its effectiveness using a quasi‐experimental design. The primary objective is to enhance students' self‐efficacy, intrinsic motivation, and higher‐order thinking skills by integrating these scaffolds into GenAI‐supported learning environments. Methods The participants were 85 university students, including freshmen, sophomores, and juniors from the School of Arts and Design. The experiment involved three groups of students: the control group, the GenAI group, and the GenAI with scaffolding group. Pre‐ and post‐tests and linear regression were conducted to assess the development of the students across the three groups. Results and Conclusions Within‐group analyses revealed that traditional learning methods without GenAI did not result in significant improvements, whereas the GenAI group exhibited statistically significant gains in intrinsic motivation, self‐efficacy, creativity, and higher‐order thinking skills. The GenAI with scaffolding group also exhibited significant improvements in critical thinking, computational thinking, and self‐efficacy. Between‐group analyses indicated that the presence of effective scaffolding (GenAI with scaffolding group) had a more positive impact on students' computational thinking and self‐efficacy compared with the absence of effective scaffolding (GenAI group). These findings indicated that while GenAI was effective in improving educational outcomes, the incorporation of SDT‐aligned scaffolding enhanced these benefits even further, which ultimately resulted in better and more enduring learning gains.