Engaged Immersive Learning: An Environment-Driven Framework for Higher Education Integrating Multi-Stakeholder Collaboration, Generative AI, and Practice-Based Assessment

Atsushi Yoshikawa

This paper proposes Engaged Immersive Learning (EIL), a new framework designed to address the passivity, lack of context, and one-way communication that often characterize lecture centered higher education. The argument begins by examining the contributions and limitations of Problem/Project Based Learning (PBL) and STEM/STEAM education. While these approaches have enhanced self directed learning and creativity through problem based inquiry and interdisciplinary collaboration, they retain structural challenges: tasks are frequently designed within the classroom, activities tend to remain short lived, assessment relies on faculty defined institutional standards, and multi stakeholder collaboration often remains superficial. Drawing on Kolb’s experiential learning, Mezirow’s transformative learning, Akpan’s social constructivism, and Lave and Wenger’s concepts of Legitimate Peripheral Participation (LPP) and Communities of Practice (CoP), the study proposes design principles for EIL. It further introduces Generative AI (GenAI) as a catalyst that supports, rather than replaces, human collaboration. EIL consists of four elements: (1) an environment driven learning space in which students are immersed for extended periods in real world contexts where multiple stakeholders (e.g., corporations, government, citizens) interact; (2) a participation trajectory in which students’ roles and responsibilities expand gradually in line with LPP; (3) a dialogue design that positions GenAI as a “buddy” that sometimes errs but offers heterogeneous perspectives; and (4) practice based assessment structured around external outcomes, stakeholder perspectives, and transformative change. Case studies from a single university illustrate that EIL can generate outcomes beyond the classroom—such as international conference presentations and forms of societal implementation—enabling employers and public officials to evaluate students in terms of “would I want to work with this person?”, and treat identity shifts as explicit learning outcomes. The paper also identifies remaining challenges concerning sample size and duration, the reliability and validity of assessment rubrics, equity of access, ethics and governance in GenAI integration, and the tension between scalability and faculty workload. EIL is therefore positioned not as a finished model but as a set of design principles to be adapted to the specific contexts of different universities and regions.

Full Text