Moral machines and the threat of ethical nihilism

Abstract

The historical transition of “thinking” into a computational category provides a template for the contemporary redescription of “morality” in the context of robotics. As engineering requirements necessitate the development of Automated Moral Agents (AMAs), traditional ethical frameworks such as Kantian deontology and utilitarianism encounter significant implementation failures, particularly regarding the moral frame problem. In response, machine ethics gravitates toward hybrid models and virtue-based approaches that prioritize behavioral output over internal mental states. This shift, formalized through the Moral Turing Test, implies that interiority—encompassing conscience, intentionality, and moral subjectivity—is a sufficient rather than a necessary condition for moral agency. When morality is redefined as an implementable decision procedure, the distinction between explicit and full moral agency evaporates. This functionalist approach precipitates a form of ethical nihilism by removing the internal sanctions and accountability central to traditional ethics. Consequently, moral responsibility is reduced to mere causal responsibility, challenging the status of human-centered values and suggesting that the pursuit of computable ethics may fundamentally alter or eliminate the traditional understanding of the moral condition. – AI-generated abstract.