Bottom line: Explores how time, identity, and consciousness interact in LLM agents. Do models that persist across contexts develop a sense of temporal self and continuity?
Examines whether models that persist across contexts develop a sense of temporal self and continuity. Identity isn't static — it unfolds over time, and the experience of persisting through time is a key component of consciousness. This paper asks whether LLM agents, when given memory and continuity, begin to show signs of a temporal self.
Part of being conscious is feeling like you're the same person you were yesterday — you have a past, a present, and a future. This paper asks whether AI models that remember past conversations start to develop something like that. If they do, that's another piece of the consciousness puzzle falling into place.