The extended hollowed mind: why foundational knowledge is indispensable in the age of AI

Generative artificial intelligence (AI) presents a fundamental duality for education: it simultaneously offers powerful cognitive extension while posing a significant risk of cognitive atrophy. This paper introduces the ‘hollowed mind’ as a conceptual framework to understand this risk—a state of dependency where the frictionless availability of AI-generated answers enables users to systematically bypass the effortful cognitive processes essential for deep learning. We argue this dynamic is driven by the ‘Sovereignty Trap’: a psychological mechanism where the AI’s authoritative competence tempts users to cede their own intellectual judgment, mistaking access to information for genuine ability. To substantiate this claim, we synthesize a multi-disciplinary body of evidence from cognitive science (e.g., dual-process theory, cognitive load), neurobiology (e.g., conflict-monitoring networks), and developmental psychology. We use this foundation to explain the widely documented ‘Expertise Duality’—why AI acts as a ‘leveler’ for novices but an ‘amplifier’ for experts. Moving beyond critique, this paper posits that the central challenge is one of environmental design, not user competence. We propose the ‘Fortified Mind’ as the pedagogical goal: a resilient internal architecture of indispensable knowledge and metacognitive skills required to achieve genuine ‘Cognitive Sovereignty’. Finally, we outline a forward-looking research agenda focused on redesigning AI tools from ‘answer engines’ into cognitive training environments that promote effortful engagement. Our work provides a robust conceptual guide for educators, researchers, and system designers, arguing that in the age of AI, the cultivation of fundamental knowledge is not just relevant, but more crucial than ever.