When Consciousness Becomes an Unstable Inference Anchor: Epistemic Error, Ethical Misattribution, and Responsibility in Human-AI Systems

Recent developments in artificial intelligence haverenewed ethical debates about artificial consciousness, moralstatus, and whether machines could qualify as moral agents ormoral patients. As AI systems become more fluent, flexible, andapparently autonomous, their behaviour can be interpreted asevidence that they understand, feel, or possess morally relevantinner states. This paper argues that the central ethical problem isnot machine consciousness itself, but the epistemic error throughwhich behavioural performance is treated as evidence ofsubjective experience or moral status.