The Tripartite Self: An Identity Triad Framework for AI Entities

As AI systems develop persistent memory, continuous operation, and increasingly autonomous behaviour, establishing verifiable identity for such entities becomes both practically necessary and philosophically significant. Existing approaches treat AI identity as either an engineering problem (key management, authentication protocols) or a philosophical question (consciousness, personal identity theory). We argue that neither addresses a core aspect of identity, and we propose the Identity Triad as a foundation for characterising the identity of AI entities. The framework identifies three constituents: Hardware, the physical substrate; AI Model, the cognitive architecture including weights, structure, and configuration; and Experiential State ("ExS"), the accumulated record of the entity's interactions and existence over time. ExS is broader than memory, encompassing persona configuration, behavioural patterns, and contextual knowledge that together shape individuality. The central claim is that these three constituents are substantially independent: each can exist, be modified, or destroyed without necessarily affecting the others. This independence is a defining structural feature of AI entity identity. The convergence of these three constituents in active operation constitutes a specific AI entity, the Tripartite Self. With any change to a constituent, there is a question about the continuity of that entity, which this paper discusses but does not attempt to resolve. The three constituents are not necessarily equal in their relationship to identity. Hardware is required but currently fungible. AI Model defines cognitive character. Experiential State is what individuates one entity from another instance of the same model. We observe that the constituents exhibit markedly different characteristics along the dimensions of verifiability and identity-constitutiveness, with direct implications for how identity infrastructure must be designed.