By changing everything we know about artificial intelligence (AI), the ways in which AI changes everything will be more plausible to explore. Arguments concerning AI as a potential threat are based on two taken-for-granted assumptions, namely, that AI exists as a separate category of intelligence, different to “natural” intelligence, and that intelligence is an inherent property of separable entities, such as humans or robots. Such arguments have given rise to ethical debates and media commentary concerned with AI, often quite extrapolating, followed by catastrophic scenarios. However, several discussions in the philosophy of social science (as well as in theoretical approaches to synthetic biology and cybernetics) have suggested (a) that the distinctions between “natural”/“human” and “artificial”/“nonhuman” are fallible, and (b) that intelligence should most likely be conceived as an environmental/systemic property or phenomenon – a shared cognition. In an attempt to import these discussions within the context of the socio-ethical implications of AI, this paper deconstructs the components of the term AI by focusing firstly on the invalidation of the term “artificial” and secondly on “intelligence.” By paraphrasing Lacan’s dictum that “the woman does not exist” as in relation to the man, this paper proposes that AI does not exist as in relation to a natural intelligence or in relation to non-intelligent entities. By this double, apparently simple, lesson learned from a re-examination of AI’s characteristics, a number of questions are raised, concerning the co-production of morality in meshed human/robotic societies, as well as a tentative agenda for future empirical investigations.