Beyond Irreplaceability: Educational Value as Self–World Transformation in the Age of Artificial Intelligence

Abstract

Generative artificial intelligence (AI) intensifies an old question about the purposes of education: why should people cultivate capacities that machines can exercise more efficiently? A prominent response defends education by identifying distinctively human abilities that AI cannot replace. This paper challenges the normative structure of that response. The fact that a task can be delegated does not settle whether a person's education is valuable; task substitutability and educational worth are different kinds of judgement. Drawing critically on Bildung, Deweyan experience, world-centred education, and theories of relationality, the paper develops a relational-transformative account of educational value. On this account, education matters insofar as it enables ethically and epistemically defensible changes in how persons encounter, interpret, and respond to themselves, others, and a shared world. Rather than treating any transformation as educationally good, the proposed framework applies non-negotiable safeguards against manipulation and serious harm, followed by three interacting criteria: epistemic answerability, reflective appropriation, and ethical responsiveness. Four contrastive cases demonstrate that AI-assisted learning may be educationally significant despite extensive delegation, whereas unaided cognitive performance need not be. The paper addresses objections concerning skill acquisition, indoctrination, pluralism, and the alleged indispensability of human cognitive labour. It concludes that AI neither abolishes nor automatically secures the worth of education: its significance depends on the forms of self–world relationship that educational practices make possible.