New Constellations of PBL in the LLM Era: Liquid Learning, Transfer, and Adaptive Expertise

Mini Track Chairs: Paul Pierce, assistant professor Informatics

Large language models (LLMs) and allied forms of generative and multimodal AI are reshaping how learners inquire, collaborate, represent knowledge, and are assessed. These developments raise a foundational question for higher education: how must problem-based learning (PBL) evolve in the age of LLMs?

Building on Savin-Baden’s notion of new constellations of PBL and contemporary calls for liquid learning—where knowledge, practices, and identities travel across contexts—this mini-track invites contributions that rethink PBL for supercomplex conditions with AI in the loop. We are particularly interested in work that prototypes agile, design-based, collaborative, and socially transformative forms of PBL in which AI augments rather than replaces human judgment, imagination, and sense-making.

LLMs are not framed here as automated tutors or answer engines, but as co-inquirers, representational tools, and reflective counterparts that can support hypothesis generation, multimodal reasoning, boundary-crossing inquiry, and collective meaning-making. Against narratives of efficiency and substitution, we foreground adaptive expertise and robust transfer—learners’ capacity to flexibly apply, reframe, and critique knowledge in novel and uncertain situations.

Recommended topics

The mini-track adopts a learning-ecologies perspective, situating PBL across interconnected micro-, meso-, and macro-systems, such as courses, programmes, institutions, professional practices, and communities. From this stance, PBL with AI is not confined to isolated classroom interventions, but unfolds across learning trajectories, infrastructures, and sociotechnical arrangements. We encourage contributions that attend explicitly to:
• Transfer across contexts and learning ecologies, rather than short-term task performance
• Adaptive expertise over routine efficiency or procedural fluency
• Ethical, sustainability, and equity considerations, including issues of power, authorship, surveillance, and environmental cost
• Student subjectification and agency in supercomplex, AI-saturated educational environments
Inspired by Barnett’s notion of feasible utopias, we especially welcome designs and analyses that stretch what universities could become while remaining grounded in educational practice. Such contributions may articulate alternative futures for PBL that are ambitious yet practicable, critical yet constructive, and attentive to the societal implications of AI-enabled education.