This paper addresses the challenge that the prospective emergence of general artificial intelligence might pose for the Thomistic version of Roman Catholic anthropology, according to which the immaterial nature of the soul is an important basis for human dignity and uniqueness. The Thomistic tradition contends that the intellect's immaterial nature can be known philosophically by reference to the intellect's powers, and especially the power to cognize universal essences. This kind of cognition presupposes that the intellect is able to receive the substantial forms of the things cognized without becoming those very things. Thomists hold that such "intentional" assimilation of forms is metaphysically incompatible with materiality. Now, to the extent that AI agents will become able to outperform humans intellectually in all or most domains, it will be increasingly difficult to maintain that AI agents lack the ability to cognize universals. Assuming, for example, that AI models will at some point be capable of producing philosophical texts that human philosophers take to be indicative of great philosophical insight (a kind of philosophical Turing test), any claim to the effect that humans have philosophical capacities that AI-agents necessarily must lack will have to rest entirely on introspective experience and theoretical assumptions. In this scenario, the Thomistic position is vulnerable to Wittgensteinian objections based on the logical relationship between mental phenomena and behavior (e.g. the private language argument). In light of the Thomistic tradition's emphasis on the harmony between faith and reason, moreover, it would be problematic for Thomists to adopt a purely fideistic stance to the existence of an immaterial soul. This paper will discuss what intellectual options Thomists might have if AI-agents were to pass a philosophical Turing-test of the most advanced kind, and what the implications of such a scenario could be for Thomistic anthropology.