On a New Ethics of AI and Moral Progress

The “new ethics” of artificial intelligence proposed by M. Gabriel is critically evaluated. It is argued that, unlike human intelligence, artificial intelligence (AI) is devoid of normative dimension, or, equivalently, of sensitivity to context. Gabriel’s view conflicts with J. Benoist’s contextual realist approach to ethics and T. Williamson’s moral realism, according to which it is not principles that are primary but moral perception in context,
paradigmatic examples of moral knowledge. The approaches of Gabriel, D. Andler, L. Floridi,    Russell to AI are considered and compared. It is proposed to adopt Andler’s principle of moderation. It is argued that AI systems imitate intelligence, agency, autonomy, ethics. A realistic conception of AI is contrasted with its idealistic conception.