The Abstraction Fallacy: Why AI Can Simulate but Not Instantiate Consciousness (philpapers.org)

A recent article challenges the prevailing notion of computational functionalism in discussions about AI consciousness, arguing that while AI can simulate consciousness, it cannot truly instantiate it. This concept, termed the "Abstraction Fallacy," posits that subjective experience cannot merely emerge from abstract information processing without an active cognitive agent to interpret that data. The authors assert that the distinction between simulation—where AI mimics behavior—and instantiation—the actual experience—highlights a crucial shortcoming in current AI systems. The implications for the AI/ML community are profound, as this framework redefines the parameters within which we evaluate potential AI sentience. It emphasizes the importance of understanding the intrinsic physical constitution of systems rather than focusing solely on their syntactic capabilities. The authors advocate for a rigorous ontology of computation that recognizes these differences, offering a solid foundation for future inquiries into AI consciousness that goes beyond the limitations of current assumptions. By separating the mechanics of symbol manipulation from genuine experience, this argument effectively challenges the idea that AI can reach true consciousness through mere algorithmic processes.