Description
AI in Education has evolved from earlier computer based systems to data driven, web based, and, more recently, generative AI environments, expanding the possibilities for formative assessment while also raising questions about transparency, ethics, and educational quality [1,2,3]. Following a PRISMA-ScR scoping review approach [4], this ongoing study examines the emerging landscape of AI enabled formative assessment in higher education, with particular attention to the tools and systems reported, their integration into teaching and learning workflows, and their main technical, pedagogical, and governance features. The review has been conducted through a machine learning assisted workflow using ASReview 2.2 [5], an open source active learning environment designed to support efficient and transparent screening while keeping the researcher in the loop. The current synthesis is based on a predefined cut off at the top 30 records in the ASReview ranking, and this flash talk presents preliminary findings from that evidence base.
The current evidence base comprises 30 empirical studies across a range of higher education contexts, including writing, programming, engineering, online learning, doctoral education, and medical education. Most studies are pilots or early implementations rather than mature systems adopted across entire institutions. Preliminary findings suggest that AI tools for formative assessment are most effective when they provide timely, structured, and revisable support for student learning while preserving meaningful human oversight. Reported systems include tools aligned with rubrics, retrieval based explainers, GPT generated feedback, systems designed to support teachers, and locally hosted platforms developed to better protect privacy. Overall, the evidence supports cautious optimism. AI may improve scalability, immediacy, engagement, and, in some cases, learning or revision quality. At the same time, robust evidence remains limited with regard to long term outcomes, transferability across institutions, explainability, governance, doctoral education, and inclusive design for diverse learners, including students with specific learning disorders, disabilities, or language and cultural barriers.
References
-
Chen L, et al. Artificial intelligence in education: A review. IEEE Access. 2020;8:75264-75278. https://doi.org/10.1109/ACCESS.2020.2988510
-
Bahroun Z, et al. Transforming education: A comprehensive review of generative artificial intelligence in educational settings through bibliometric and content analysis. Sustainability. 2023;15(17):12983. https://doi.org/10.3390/su151712983
-
Ahmad K, et al. Data-driven artificial intelligence in education: A comprehensive review. IEEE Trans Learn Technol. 2024;17:12-31. https://doi.org/10.1109/TLT.2023.3314610
-
Tricco AC, et al. PRISMA extension for scoping reviews (PRISMA-ScR): checklist and explanation. Ann Intern Med. 2018;169(7):467-473. https://doi.org/10.7326/M18-0850
-
ASReview LAB developers. ASReview LAB: a tool for AI-assisted systematic reviews. Version 2.2. Zenodo. 2025. https://doi.org/10.5281/zenodo.17817658
Link to the research group: https://cqm.unicam.it