인공지능의 도덕적 행위자로서의 가능성: 쉬운 문제와 어려운 문제

초록

I discuss if AI can be regarded as moral agent by dividing the issues into two categories: easy problems and difficult problems. I illustrate them concepts using the case of autonomous vehicles. Then I argue that in order to tackle the easy problems, we need to ‘extend’ socially accepted ethical intuitions and legal frameworks. I explain how to achieve this by following the historical analysis of legal person concept. I also claim that the resolution of ‘difficult problems’ can be obtained through our continous resolution of a number of the related easy problems. Finally, I consider a potential critique that my approach is impossible to implement as the two kinds of problems are essentially interlinked, and reply to it by pointing out the constructive and social aspects of my approach.

키워드