A interesting instance of social and individual rationality contradicting each other is the question of how to program self-driving cars to react in the case of a looming accident. Ignoring the lurid headline the article on this social dilemma is quite interesting. Apparently, early research by Toulouse psychologist Jean-Francois Bonnefon indicates the majority of people agreeing that cars should be programmed to choose the alternative which kills the least amount of people in expectation. Unsurprisingly, this ceases to hold true if this increases the probability of the interviewee being killed in the accident.