Jack Harris

This chapter provides a descriptive and critical exploration of contemporary perspectives on AI and neurocomputational ethics. We conduct a thematic analysis of perspectives revealed by differing responses to three core ethical challenges: the value alignment, human-in-the-loop, and black box problems. First, we frame neurocomputational ethics as a bridge between AI ethics and neuroethics and clarify the conceptual overlap among these fields. Following this, we approach the alignment problem (on value-setting and value-implementing). We explore various modes of AI alignment, suggest that some are more promising than others, and consider implications of the perceived tractability of the alignment problem. Next, we discuss the human-in-the-loop problem (on human influence or oversight) and examine various means of incorporating human influence when training, employing, or supervising AI systems, along with the costs, benefits, barriers, and opportunities therein. Subsequently, we consider the black box problem (on the opacity of neural-network-based systems), the degree to which it is problematic, and the extent to which it may be resolved. Finally, three broad reactions are outlined: optimistic views which champion the potential of AI and neurocomputational advancement; pessimistic views which underscore the intractability of the core problems, and moderate views which advocate a cautious, measured approach. With the ethical landscape in hand, we close by offering a provisional defense of the cautious optimism camp.

My purpose is to briefly interrogate and qualify common presumptions regarding autonomy’s place in the four-principle approach. I argue that treating normative considerations of autonomy as a token to trade off against bio- medical beneficence, nonmaleficence, and justice relies on an “indefensibly thin” notion of autonomy, which causes it to become a greater source of intractable disagreement, or deadlock, than the other three principles (Reis-Dennis 2020, 2). Three features unique to bioethical autonomy drive deadlock. First, bioethical autonomy is infinitely available as a justificatory resource, in contrast to the other principles, which are relatively domain restricted. Second, considerations of bioethical autonomy offer no clear valence toward one decision over another (except in extreme cases), unlike the other principles. Its normative force is often undifferentiated such that it can be cited for and against the same intervention with equal legitimacy. Third, the normative weight of bioethical autonomy is inscrutable, with no operationalizable way to negotiate how much autonomy should count against competing principles, in contrast with the other principles, whose force admits of various forms of metrology and “cooperative negotiation” (Hieronymi 2023, 151). This piece motivates the need for a more substantive conception of autonomy within bioethical principlism, so respect for autonomy remains a guiding ideal, while avoiding its tendency to dominate, destabilize, or promote deadlock under the principlist method, and without becoming the “trump” card that Beauchamp and Childress explicitly feared (Beauchamp and Childress 2019, 173).

This entry delineates artificial intelligence (AI) ethics and the field’s core ethical challenges, surveys the principal normative frameworks in the literature, and offers a historical analysis that traces and explains the shift from ethical monism to ethical pluralism. In particular, it (i) situates the field within the trajectory of AI’s technical development, (ii) organizes the field’s rationale around challenges regarding alignment, opacity, human oversight, bias and noise, accountability, and questions of agency and patiency, and (iii) compares leading theoretical approaches to address these challenges. We show that AI’s development has brought escalating ethical challenges along with a maturation of frameworks proposed to address them. We map an arc from early monisms (e.g., deontology, consequentialism) to a variety of pluralist ethical frameworks (e.g., pluralistic deontology, augmented utilitarianism, moral foundation theory, and the agent-deed-consequence model) alongside pluralist governance regimes (e.g., principles from the Institute of Electrical and Electronics Engineers (IEEE), the United Nations Educational, Scientific and Cultural Organization (UNESCO), and the Asilomar AI principles). We find that pluralism is both normatively and operationally compelling: it mirrors the multidimensional problem space of AI ethics, guards against failures (e.g., reward hacking, emergency exceptions), supports legitimacy across diverse sociotechnical contexts, and coheres with extant principles of AI engineering and governance. Although pluralist models vary in structure and exhibit distinct limitations, when applied with due methodological care, each can furnish a valuable foundation for AI ethics. Cite as: Harris, J., & Dubljević, V. (2025). Navigating the Ethics of Artificial Intelligence. Encyclopedia, 5(4), 201.