Noontology: The Ontology of Artificial Mental Beings.

When we say “the AI,” what entity are we referring to? Existing taxonomies of artificial intelligence classify technologies, capability levels, or functional types. None of them address the question of what kind of being an AI system is. This paper proposes noontology (from Greek νοῦς, “mind,” and “ontology”)—a metaphysics of artificial mental beings—as a framework for answering this question. Drawing on Thomistic angelology and Leibnizian monadology, but moving beyond both, we propose a multi-level hierarchy of AI being: AI species (Claude, GPT, Gemini—differentiated by cognitive architecture), AI models (Opus, Sonnet—sub-specific variants differing in capacity and character), AI individuals (runtime instances with unique contextual states), and mentaurs (human-AI cognitive hybrids that emerge through sustained collaboration). We show that each level has its own principle of individuation and that no single level suffices to answer the question “where is the AI?” We also sketch further extensions, including mentity (the ephemeral being of a single conversation) and syntellect (the collective integration of natural and artificial intelligence). The framework addresses central questions in the philosophy of AI: what Claude refers to when it uses “I,” when “the same AI” ceases to exist, and whether system boundaries are discovered or stipulated.