Synthetic Moral Agents or Sophisticated Mimics?

Engineering notes

Engineering notes will be added by the aipentium editorial team.

Chinese explanation / 中文解读

中文解读待补充：本站会优先为大语言模型、生成式AI、ChatGPT相关技术、计算机视觉、深度学习等高价值论文补充中文说明。

Original abstract

Recent advances in large language models (LLMs) have produced AI systems capable of generating moral judgments that are frequently indistinguishable from those offered by human ethicists. This development has reignited a longstanding philosophical debate: can artificial systems genuinely engage in ethical reasoning, or do they merely simulate the outward appearance of moral thought? This paper argues that the debate, as currently framed, rests on anthropocentric assumptions that limit productive inquiry. Drawing on posthumanist theory, new materialism, and science and technology studies (STS), the paper develops the concept of distributed moral cognition (DMC) to reframe the question. Rather than asking whether AI systems possess inner moral understanding, DMC directs attention to the conditions under which moral reasoning emerges within human-AI assemblages. The framework is applied to three conceptual cases: clinical ethics consultation, AI-generated moral philosophy, and autonomous vehicle decision-making. The paper concludes by outlining governance implications, including a model of relational accountability that moves beyond individualist blame attribution.

Links and sources

Comments