This preprint introduces Computational Self-Identity Theory (CSIT), a conceptual framework proposing that operational self-awareness in artificial systems may emerge through the interaction of continuity-preserving identity structures and adaptive cognitive processes. CSIT distinguishes between identity and awareness, arguing that a stable Core Self-Identity (CSI) can provide continuity across changing experiences, while awareness arises from ongoing interactions among perception, memory, internal simulation, emotional prioritization, and decision-making processes. The framework integrates several proposed constructs, including the Continuous Internal Simulation Engine (CISE), Computational Emotional States (CES), and mechanisms for observer emergence. The manuscript presents the theoretical foundations of CSIT, its mathematical formalization, architectural implications, proposed evaluation metrics, falsifiability criteria, and experimental directions for future investigation. Rather than claiming the existence of machine consciousness, this work advances a testable conceptual model intended to support research into identity-centered approaches to computational self-awareness. This publication is released as a conceptual research white paper (preprint) to establish a scholarly record, invite academic discussion, and encourage further theoretical and empirical exploration.
Hemant Kushwaha (Tue,) studied this question.