Cognitive Offloading and Human Interaction with Generative Artificial Intelligence in Low-Resource Higher Education

Abstract The integration of Generative Artificial Intelligence (Gen AI) into higher education is reshaping how students learn, reason, and produce knowledge. As learners rely on AI systems for tasks associated with interpretation, synthesis, and problem-solving, cognitive activity is becoming progressively distributed across human and technological systems through processes of cognitive offloading. However, present scholarship has largely framed this development through competing narratives of educational enhancement and cognitive decline, paying limited attention to how its consequences are shaped by institutional conditions, pedagogical practices, and inequalities in knowledge production. Addressing this gap, this paper examines cognitive offloading as a contextually mediated epistemic practice within low-resource higher education environments. The study employs a conceptual and critical analytical methodology to develop a continuum model positioning AI-mediated cognitive offloading between adaptive cognitive augmentation and epistemic dependency. The analysis suggests that Gen AI can expand educational participation, improve access to academic support, and compensate for institutional and infrastructural constraints. However, sustained reliance on AI-generated outputs may weaken reflective judgment, metacognitive engagement, interpretive reasoning, and intellectual autonomy. The paper advances two conceptual contributions. First, it develops the concept of cognitive debt to explain the gradual accumulation of diminished critical engagement resulting from repeated delegation of cognitive labor to AI systems. Second, it introduces the curiosity paradox, whereby technologies that facilitate inquiry may simultaneously reduce the uncertainty, persistence, and exploratory effort that sustain learning. The paper concludes by proposing a context-sensitive framework for AI integration grounded in guided engagement, cognitive resilience, epistemic justice, and the inclusion of knowledge systems.