Awad, Edmond (2017). Moral Machine: Perception of Moral Judgment Made by Machines (Master's Thesis).

Abstract

While technological development of vehicular autonomy has been progressing rapidly, a parallel discussion has emerged with regard to the moral implications of a future wherein people hand over to autonomous machines the controls to a mode of transportation. These discussions have entered a new phase with the US Department of Transportation (DoT) releasing a 15-point policy that requires manufacturers to explain how their AVs will handle “ethical considerations." However, there is a huge gap in our understanding of the ethical perception of AI, as there have been few large-scale empirical studies on human moral perception of outcomes to autonomous vehicle moral dilemmas. Additionally, public engagement is a very important piece of the puzzle, especially given the emotional salience of traffic accidents. With that in mind, I co-developed the Moral Machine. Moral Machine is a platform for gathering a human perspective on moral decisions made by machine intelligence, such as AVs. The website went viral, and got covered in various media outlets. This website has also been a valuable data collection tool, allowing us to collect the largest dataset on AI ethics ever collected (with 30 million decisions by over 3 million visitors, so far). This thesis will introduce Moral Machine as a data-gathering platform. Moreover, insights about the human perception of the different routes to full automation will be covered in the thesis, with the data collected through other online platforms.