The Virtuous Machine - Old Ethics for New Technology?

Virtue ethics offers a practical framework for AI ethics: Rather than relying solely on rule-based approaches, the paper argues that Aristotelian virtue ethics—which emphasizes learning from experience and developing good character traits—is well-suited to modern AI systems because it aligns with how these systems actually learn.

AI systems could learn morality through imitation of exemplars: The paper proposes that the value alignment problem (ensuring AI systems pursue human-compatible goals) could be addressed by training AI systems to learn from "moral exemplars"—essentially teaching machines to imitate the behavior of ethically sound role models, similar to how humans develop moral understanding.

Virtues like temperance could address the "control problem": The paper suggests that if AI systems were developed with virtues such as temperance and "friendship to humans," they would naturally lack the desire for unlimited self-improvement, potentially solving concerns about uncontrollable AI systems that continuously seek to enhance themselves beyond human oversight.

Autonomy and proximity to humans create new ethical urgency: As autonomous machines become more capable and work closer to people—in autonomous vehicles, service robots, and personal assistants—their actions take on genuine moral significance, making it essential to build ethical considerations into their design from the start.