The central tension explored in this article isn’t simply about whether AI *can* be virtuous, but rather *how* we understand virtue itself, and whether the very category of “virtue” is compatible with the architecture of current AI systems. The article presents a layered critique, beginning with the relatively optimistic framing of Anthropic’s Claude – an attempt to create an AI agent aligned with human values – and then dismantling that framing through the lens of Thomistic virtue ethics. Fathe...