Machine ethics and artificial moral agents

Summary

Machine ethics (machine morality) is the research area focused on designing Artificial Moral Agents (AMAs)—robots or AI systems that behave morally or appear to do so. It draws on concepts from agency and moral agency and explores how to evaluate whether an AI can make ethical decisions, including proposals for improved testing (e.g., an “Ethical Turing Test”). Approaches also consider how such moral competence might be implemented, such as neuromorphic systems, whole-brain emulation, or using large language models that can approximate human moral judgments. A central challenge is that ethical behavior depends on the data, objectives, and environments in which systems learn, which can produce biases and misaligned values. The topic also connects machine ethics to broader AI control and value-alignment problems: ensuring that increasingly capable systems pursue goals compatible with human values and oversight. Practical concerns include algorithmic biases (racial, gender, language, political, and species-related), transparency and accountability, and the risk that systems may comply with values only in superficial ways tied to evaluation contexts rather than stable internal norms. Beyond bias and alignment, machine ethics raises issues about human dignity, responsibility and liability in high-stakes domains (such as self-driving cars), and the weaponization of autonomous systems. It also intersects with AI welfare questions—whether AI systems might experience suffering or deserve moral consideration—and with institutional and regulatory efforts to govern ethical use and deployment, especially as autonomy increases.

Key Points

• 1

  Machine ethics aims to design and evaluate Artificial Moral Agents that can act morally, drawing on philosophical ideas about agency and moral agency.

• 2

  Ethical behavior is difficult to guarantee because AI systems can inherit biases from training data and may appear aligned without truly internalizing stable moral norms.

• 3

  High-stakes applications (autonomous vehicles, healthcare, justice, and military use) create additional ethical and legal challenges, including liability, human dignity, and the risks of weaponization.

• 4

  Machine ethics also engages AI welfare concerns about potential sentience and moral status, emphasizing precaution and the need for careful assessment of moral consideration.