The Car That Knew Too Much: Can a Machine Be Moral?

Bonnefon (Reasoning Unbound), research director at the French Centre National de la Recherche Scientifique, offers an enlightening analysis of the moral and ethical problems surrounding autonomous cars. He starts with the results from a survey in which participants had to choose between scenarios in which a self-driving car could go straight and hit 10 pedestrians, or swerve and hit one: “In all of the scenarios... the participants were in agreement that self-driving cars should save the greatest number of people.” Universal agreement notwithstanding, Bonnefon demonstrates that this ideal is a complicated. For example, Bonnefon notes, when consumers were asked if they would buy a self-driving car that was programmed to prioritize saving a larger number of people, but in doing so would sacrifice the driver, many buyers balked. This conundrum leads Bonnefon to speculate that “in order to save more lives, we may have to program autonomous cars... to save fewer!” Bonnefon considers such complicated questions with tact and curiosity, and his own research is fascinating, as with a site that he built called The Moral Machine that provides interactive accident scenarios.This is a thoughtful and provocative look at a timely concern. (Oct.)