Artificial Intelligence in Practice-Based and Creative Learning Pedagogies in Higher Education: A Systematic Review of Instructional Designs and Learning Outcomes

Abstract

This systematic review examines how artificial intelligence (AI), including generative and non-generative systems, is integrated into practice-based and creative learning in higher education, and how strongly the reported outcomes are supported by the underlying evidence. Searches of Scopus, Web of Science, and ERIC identified 166 records. After duplicate removal, title/abstract screening, retrieval, and full-text assessment, 38 studies were retained for synthesis after one conference paper that conflicted with the stated eligibility criteria was removed. The included literature was concentrated in architecture/interior design, art/visual arts, and graphic/visual design. Studio-based pedagogy was the most frequent primary format (14/38, 36.8%), followed by project-based learning (11/38, 28.9%). AI most often supported ideation and visualization (23/38, 60.5%), although 15 studies embedded AI across multiple stages of the learning or design cycle. Twenty-eight studies (73.7%) were coded as positive in direction, four (10.5%) as mixed-to-positive, and six (15.8%) as mixed. These directions should not be interpreted as equivalent causal effects: eight studies relied primarily on objective tests/performance, eight on artefact/expert evaluation, eight on self-report, and the remainder on mixed or process evidence. Methodological appraisal using the Mixed Methods Appraisal Tool (MMAT) identified eight studies with relatively higher confidence, 23 with moderate confidence, and seven with lower confidence within this review corpus. The evidence therefore supports a cautious conclusion: AI commonly expands ideation, visualization, feedback, and workflow options, while stronger claims about improved creativity or learning depend on controlled designs, objective/artefact-based assessment, and longer-term evidence.    Received: 22 April 2026 / Accepted: 13  August  2026 / Published: September  2026