Ethical Rationality in AI

Taking its starting point in a discussion of the concept of intelligence, the chapter develops a philosophical understanding of ethical rationality and discusses its role and implications for two ethical problems within AI: Firstly, the so-called “black box problem,” which is widely discussed in the AI community, and secondly, another more complex one which will be addressed as the “Tin Man problem.” The first problem has to do with opacity, bias, and explainability in the design and development of advanced machine learning systems, such as artificial neural networks, whereas the second problem is more directly associated with the prospect for humans and AI of becoming full ethical agents. Based on Aristotelian virtue ethics, it will be argued that intelligence in human and artificial forms should approximate ethical rationality, which entails a well-balanced synthesis of reason and emotion.

The Bright Line of Ethical Agency

In his article The Nature, Importance, and Difficulty of Machine Ethics, James H. Moor distinguishes two lines of argument for those who wish to draw a “bright line” between full ethical agents, such as human beings, and “weaker” ethical agents, such as machines whose actions have significant moral ramifications. The first line of argument is that only full ethical agents are agents at all. The second is that no machine could have the presumed features necessary for ethical agency. This paper shows why Moor is mistaken in his refutation of the first line of argument; it also makes a positive case that “weaker” ethical agents are not agents at all. This positive case, however, allows Moor’s rejection of the second line of argument to stand: allowing that there could be moral machines, but that these machines would have to be full moral agents and not merely something that models moral behavior or can be used in morally charged ways.