Agentic AI - defined by autonomous decision making, persistent contextual awareness, goal-directed behaviour, and the capacity to independently initiate instructional actions - represents a qualitative shift from earlier educational technologies. Unlike generative AI tools, intelligent tutoring systems, adaptive learning platforms, or human-in-the-loop systems, agentic AI functions not as a tool that supports human decision making but as a co-participant in learning. This distinction raises a question that existing research has not adequately addressed: how does agentic AI reconfigure the core pedagogical encounter - the interactions between students, educators, and content - and what tensions does this redistribution of agency produce?

Concepts such as human-AI co-agency and sociotechnical learning partnerships are gaining traction but remain theoretically underdeveloped. What is poorly understood is how agentic AI affects pedagogical authority, learner autonomy, epistemic trust in AI-generated feedback, and the risk of cognitive offloading and learned dependency. These tensions are central to determining whether agentic AI can be integrated into higher education in educationally responsible ways.

This Research Topic focuses on the pedagogical and experiential dimensions of agentic AI in higher education - not on institutional governance or broad digital transformation - asking how agentic AI reconfigures what happens inside the teaching and learning process itself.

We seek contributions examining: how pedagogical authority and instructional roles shift when AI takes on functions traditionally performed by educators; how students experience autonomous or semi-autonomous AI, including effects on motivation, engagement, and self-regulation; how human-AI co-agency can be theorised and empirically studied, including its tensions and asymmetries; how curriculum and assessment practices must adapt when AI actively shapes learning pathways and contributes to feedback and evaluation; how epistemic trust and cognitive dependency develop or erode over time; and how disciplinary differences shape integration across fields.

We welcome contributions that use sociomaterial, sociocognitive, constructivist, distributed cognition, self-determination, and critical pedagogy frameworks analytically - as tools for interrogating tensions and redistributions of authority, not merely as descriptive lenses. Longitudinal, interaction-focused, and process-oriented designs are especially encouraged.

Topics of interest include: pedagogical authority and role reconfiguration; student experiences with autonomous tutoring or feedback agents; human-AI co-agency models and their tensions; instructional design for agentic AI environments; assessment validity and epistemic trust; effects on autonomy, self-regulation, and cognitive dependency; disciplinary variation in agentic AI integration; comparative studies of agentic versus assistive AI; design-based research on agentic AI interventions; and critical perspectives on the redistribution of pedagogical power.

We welcome Original Research, Reviews, Systematic Reviews, Mini Reviews, Methods articles, Perspectives, and Conceptual Analyses from interdisciplinary teams bridging education, learning sciences, psychology, computer science, and discipline-specific pedagogy.

Are you interested in contributing a manuscript to this Research Topic?

Agentic Artificial Intelligence in Higher Education: Pedagogical Transformation, Student-Teacher Interactions, and Human-AI Co-Agency