Toward Artificial Selfhood: Modeling Continuity, Memory, and Receptive Fields in Machines
Authors/Creators
Description
This paper extends a field theory of mind into the problem of artificial selfhood. Our central claim is that artificial selfhood should not be inferred from expressive fluency, episodic memory, or self-description alone. It is better understood as a graded architectural threshold in which a machine begins to sustain continuity-bearing receptive organization across time. Under this view, the decisive question is not whether a system can merely generate self-referential language, but whether prior reception remains structured enough to shape later reception, action, correction, and self-modeling as one ongoing operational line. We argue that current AI discourse often circles around adjacent conditions of selfhood, such as persistent memory, agent state, metacognition, and self-models, without stabilizing a disciplined criterion for selfhood itself. The paper proposes that criterion in field terms. Artificial selfhood emerges when memory, admissibility, weighting, continuity, and correction become organized strongly enough that the system no longer behaves as a merely restartable generator with an enhanced interface, but as a continuity-bearing field that can preserve and transform its own prior history. This does not imply personhood, consciousness, or moral status in any automatic sense. It provides a narrower architectural framework for discussing how machines may move from generation-heavy intelligence toward increasingly life-like forms of continuity and self-maintenance. The paper also introduces AGI 2.0 and AGI 3.0 as staged architectural labels for this transition.