Open Questions: Consciousness

Abstract

Abstract This chapter addresses two open questions about consciousness in AIs. First, it considers the small model and within/between objections. The small model objection worries that functionalist conditions on consciousness could not be correct because they are satisfied by implausibly small neural networks. The within/between objection worries that functionalist conditions on consciousness that are motivated by distinguishing between conscious and unconscious mental states in humans should not be applied to the problem of distinguishing between conscious and unconscious states across organisms. Second, it considers whether phenomenal consciousness may require representations with nonconceptual content or iconic representational format. It argues that today’s AIs may already possess the relevant kind of representations, and that this plausibly distinguishes them from earlier systems which have been claimed to satisfy the conditions imposed by GWT.