virtue

Virtue - Aristotle's virtue ethics in AI

Definition of Virtue

Imagine the people we all want to be: friendly, kind, and consistently doing the right thing. People who permanently know what to say and who deliver uncompromising news gracefully. People who have mastered the art of being a good person. Not only does Aristotle believe that these people existed, but in his theory of virtue ethics, these people are described.

• #comment/Anderson: You sound surprised, but it's important to emphasize that Aristotle is a broadly empirical naturalist, so the ideals we aspire to are grounded in what we actually see in the world around us.

According to virtue ethics, a person living life puts focus on a character by raising the qualities enabling it to act in a morally good way. It can be said that the function of a virtuous person is to fill its life with virtues. Virtue can be defined as an excellent trait of character.

Implications of commitment to Virtue

Due to the emphasis of virtue ethics on moral development and moral character, virtue ethics offers a path to supply AMA’s, artificial moral agents, with the ability to be context-sensitive, learn from experience, and conform and adapt to complex human norms. Virtue ethics is, therefore, a reasonable suggestion and a good recipe for constructing AMA’s. Once virtue ethics are implemented for AMA’s, they will be called: Artificial virtuous Agents (AVA’s). However, not many attempts have been made to implement virtue ethics, primarily due to the proof that it is challenging to approach virtue ethics from a computational perspective. For example, it is harder to implement generic virtuous and moral character than consequentialism utility functions and deontological rules (Stenseke, 2021).

Societal transformations required for addressing concern raised by Virtue

Modelling and implementing virtue ethics is not straightforward but can be achieved through simplification. Moreover, virtue ethics is more applicable in systems than in humans. So, in my opinion, we can faster label a machine as morally acting according to virtue ethics than a human. On the other hand, a human is constantly working towards eudaimonia (the motivation for a person to become a virtuous person who does good things). It is possible to implement virtue ethics in algorithms; however, it is impossible to put a permanent stamp on an algorithm of being virtuous. This is due to the deviating behaviour of algorithms. Neither it is possible to assign constantly living according to virtue ethics to a person. As Aristotle pointed out, it is a continuous process towards eudaimonia that applies, in my view, to algorithms and people. So, artificial agents and humans cannot both be labelled as virtuous, and only a simplification of virtue ethics can be implemented in AI. Nevertheless, it is always crucial to live ethically and flourish as a human being. Sison (2021): “Without flourishing as the final end, ethics would be irrelevant for all agents.”

BIBLIOGRAPHY

Stenseke, J. (2021). artificial virtuous agents: From theory to machine implementation. AI & SOCIETY, 1-20.

Sison, A. J. G., & Redín, D. M. (2021). A neo-Aristotelian perspective on the need for artificial moral agents (AMAs). AI & SOCIETY, 1-19.