91�Թ���

Could Artificial Intelligence (AI) ever become conscious? The hard problem of consciousness bears on this question. Chalmers (2010) defines the “hard problem” as an unbridgeable epistemic gap between physical and phenomenal states, which he claims entails an ontological gap and a form of ontological dualism. The hard problem challenges physicalism, a form of ontological monism, where consciousness emerges from the right physical structure. Frequently in the literature, physicalists are regarded as reductionists and skeptical of AI consciousness (see Aru et al. 2023), whereas dualists are regarded as anti-reductionists and more sympathetic (see Block 1997).

I argue that physicalism and dualism should not be conflated with reductionism and anti-reductionism respectively. To decouple these debates, I propose a novel account of non-reductive physicalism that maintains an epistemic gap while rejecting an ontological gap. My account confines ‘levels of relative fundamentality’ to scientific theories and refrains from transforming them into ontological hierarchies. Finally, I contend that it’s important to pursue non-reductive physicalism as a solution to the hard problem in order to reflect advances in AI research.