The wandering guide in the age of AI: reconceptualizing nomadic heutagogy for self-determined learning

Heutagogy positions learners as the authority over their own learning, and nomadic heutagogy, grounded in Michel Serres' philosophy, frames this authority as wandering freely through knowledge networks. Generative AI introduces a paradox to this vision: while it lowers barriers to exploration, it may also foster a gradual, often invisible slide from trust to dependence, bypassing the cognitive struggle essential to deep learning. We present a conceptual analysis of how generative AI challenges four constructs central to nomadic heutagogy: learner autonomy, knowledge co-creation, metacognition, and the facilitator role. Across all four, we find that what matters is not AI itself but how learners relate to it as a knowledge source. We identify three stances: oracle, in which the learner defers to AI as authority; tool, in which the learner uses AI instrumentally; and co-wanderer, in which the learner engages AI as a thinking partner that prompts inquiry rather than settles it. Based on this analysis, we propose design principles to guide learners toward the co-wanderer stance. Central to our framework is the concept of epistemic self-regulation, the learner's growing capacity to monitor their relationship with AI and recognize when they are shifting from trust to dependence. Rather than stepping back to enable autonomy, the guide's role becomes cultivating this capacity. We extend nomadic heutagogy to address conditions its originators could not have anticipated, offering educators a principled framework for preserving self-determined learning in an era when AI is ever-present.