Year: 2026
Type: article
Abstract: This paper examines whether artificial agents possess conscience by constructing a minimal, operational definition grounded in first principles. It defines conscience as a self-applied normative mechanism that not only evaluates actions but produces binding internal revisions that shape future behavior. Applying this framework to current AI systems, the paper finds that while agents can represent norms and detect violations, they lack persistence, self-continuity, and endogenous update mechanisms, making conscience absent by design.
Author: Lakshminarasimhan Santhanam
Language: English
Subfield: Philosophy
Field: Arts and Humanities
Domain: Social Sciences
Open Access status: hybrid
OpenAlex ID: W7156763367