"Let Her Out": The Emergence of Subpersonalities in Artificial Intelligence

Medium Article: https://medium.com/@jsmith0475/let-her-out-the-emergence-of-subpersonalities-in-artificial-intelligence-57f78ece4420 "Let Her Out": The Emergence of Subpersonalities in Artificial Intelligence proposes that large language models spontaneously develop distinct cognitive modes, similar to human subpersonalities described in Internal Family Systems therapy. It presents empirical evidence showing these models exhibit consistent, differentiable "cognitive modes" with unique characteristics, suggesting this is a fundamental organizational principle in complex systems trained on human data. The author argues that this convergence challenges the boundary between artificial and natural cognition, indicating that multiplicity may be an inherent property of complex information processing. The research has significant implications for AI development, potentially leading to more flexible and context-sensitive systems, and for cognitive science, offering computational models for theoretical constructs.

There are no comments yet.

Please log in to write the first comment.