Persons or Agents: The Ontological Stakes of Education in the Age of Artificial Intelligence

The question of education in the age of artificial intelligence is typically framed in instrumental terms: what skills to teach, what digital tools to integrate, how to prepare students for a transformed labour market. This article argues that the question is not pedagogical but ontological. Drawing on a five-axiom philosophical anthropology — person/thing distinction, irreducibility of practical wisdom, the person as an end, responsibility as constitutive, and the capacity to initiate — we show that the canonical definition of "agent" in artificial intelligence theory maps exactly onto the hyperspecialised professional produced by contemporary instrumentalised education. The correspondence is not metaphorical but ontological: the human being trained exclusively for functional performance, without tools for moral judgment or reflective self-awareness, satisfies the definition of agent point by point. We distinguish two educational models: person-formation (Model A), which cultivates the capacities identified in the axioms, and agent-production (Model B), which systematically atrophies them. Building on companion articles that establish the non-delegability of moral judgment and the role of telos in determining whether AI deployment is humanising or dehumanising, we argue that education is where the telos is formed: the disposition to design technology from the person rather than from the system. The article's most consequential claim is recursive: only persons can choose to form persons; if a generation is educated without the tools of moral judgment, the choice itself becomes impossible — not because anyone prohibits it, but because no one remains capable of making it. This article is the third in a triptych on AI and personhood. The foundational manifesto is available at https://doi.org/10.5281/zenodo.18650485.