Naming AI Counterparts: A Communicative Account of Individuation and Continuity
GeonWoo Kim · PhilPapers (PhilPapers Foundation)
When an AI system is updated, divided among agents, or moved across devices, is a returning user still dealing with the same counterpart? The same tasks and materials can accompany a change of counterpart. This conceptual article argues that naming organizes connections between encounters and particular counterparts, and that these connections can be maintained and revised in parts. It links three tasks of individuation to the criteria for counterpart continuity: specifying a present target, linking past encounters as its history, and identifying it again elsewhere. An adopted designation can be reused across changes of role, encounter, and position of address. The analysis follows what that reuse connects, alongside introductions, accounts of the past, and identification records. Constructed contrasts show a request continuing with one counterpart or passing to another, an address being corrected, and encounters gathered in a report to a third party. Variants involving renaming, corrected attribution, and technical replacement establish why these connections need not stand or fall together. A name may change while the counterpart connection holds; an attribution may be corrected while the review history remains; a replacement record may overturn an identification while the name persists. Technical evidence bears on each identity claim through the criterion in use for the counterpart concerned. The account develops historical and interactional approaches to reference within research on artificial identity, communicative anthropomorphization, and name effects. It explains which continuity a designation carries in use and offers observation questions and design examples for making that continuity intelligible.