자율주행차량이 상용화되기 전에 해결되어야 할 윤리적ㆍ법적 난제와 그 해법 - 윤리적 가이드라인과 Moral Machine Experiment를 중심으로 -

Abstract

Nowadays, lots of nations, institutions and companies have issued principles or guidelines for ethical artificial intelligence system and autonomous driving car. This paper introduces “Ethical rules for automated and connected vehicular traffic” by Ethics Commission -Automated and Connected Driving- in Federal Minister of Transport and Digital Infrastructure, June 2017 in Germany and “Ethics Guidelines for Trustworthy AI” by Independent High-Level Expert Group on Artificial Intelligence set up by the European Commission, to outline the current state of discussions on ethical guidelines for artificial intelligence systems. In conclusion, although there are some differences, the main contents of the ethical guidelines currently being presented in various countries around the world seem to be summed up in five principles: i. e., transparency, justice and fairness, non-malefic ence, responsibility and privacy, as can be seen from the research by Jobin et. al.. We could use these principles in the process of deriving ethical guidelines for AI in Korea too. However, our expectation that it will clearly provide ethical standards for what to sacrifice in an inevitable accident situation known as the so-called trolley dilemma, is not met. In this context provides a research of Rahwan et. al, so called “Moral Machine Experiment”, a good basis for why it is difficult to derive universal standards in an ethical dilemma. In other words, with regard to what kind of life will be saved in a dilemma, people make different judgments based on various factors such as country, region, religion, culture, gender, or economic conditions etc. to which they belong. Furthermore, it also expresses the problem to be solved by ethical and legal guidelines in a seemingly contradictory expression of those who will not purchase their own self-driving vehicles, even though they think that it is ethically correct to save pedestrians and to sacrifice passengers, in other words ironically to sacrifice themselves. The author wants to say, in order to establish ethical and legal guidelines for AI systems or autonomous vehicles equipped with artificial intelligence in a society where diverse ideas are mixed, the most urgent task is to check ethical preference of citizens living in Korea.