From Data to Mind: A Critical Evaluation of AGI and Machine Consciousness

Abstract

Artificial General Intelligence (AGI) — a hypothetical system capable of matching or exceeding human cognitive performance across arbitrary domains — remains one of the most consequential and contested goals in computer science. Parallel to this pursuit is the deeper philosophical question of whether such a system could possess genuine subjective experience, or machine consciousness. This paper critically evaluates the feasibility of AGI and machine consciousness by synthesizing evidence from artificial intelligence research, cognitive science, and philosophy of mind. We review dominant technical approaches to AGI — symbolic, connectionist, hybrid, and whole-brain-emulation paradigms — and assess their trajectories against benchmarks of generality, transfer learning, and reasoning. We then examine four influential theories of consciousness — Integrated Information Theory, Global Workspace Theory, Higher-Order Theories, and Predictive Processing — and analyze their implications for machine sentience. Our analysis suggests that while narrow capability gains are accelerating rapidly, the transition from sophisticated data-driven cognition to genuine general intelligence, and further to consciousness, involves conceptual and empirical gaps that current architectures do not resolve. We conclude by outlining the ethical stakes of this uncertainty and proposing directions for interdisciplinary research.