AI Consciousness Requires Validated Models of Human Consciousness

Claims about AI consciousness should be grounded in models first validated on humans, because scientific observation ultimately depends on human perceptual agreement. Without such models, the question of whether an AI is conscious lacks sufficient empirical grounding. The authors propose a human-first methodology: identify measurable phenomena linked to human consciousness, build and validate predictive models, and only then apply them to AI systems. This approach aims to transform philosophical debates into productive scientific inquiry.

Study at a glance

Abstract

Debates about AI consciousness often proceed without grounding the concept in empirically validated models. This position paper argues that meaningful claims about AI consciousness should be licensed by (and graded by confidence in) models validated on humans. Drawing on Quine's observation sentences and pragmatic philosophy of science, we argue that all scientific observation ultimately depends on human perceptual agreement, including observations about consciousness itself. Without validated human models that make testable predictions about conscious experience, the question ``Is this AI conscious?'' lacks sufficient empirical grounding necessary for scientific progress. We propose a human-first methodology: identify measurable phenomena associated with consciousness in humans, build predictive models, validate them empirically, and only then apply these models to AI systems. This approach accelerates philosophical debates into productive scientific inquiry.