Algorithms of life and death: a utilitarian approach to the ethics of self-driving cars

The thought of self-driving cars operating on our roads is no longer the stuff of science fiction. Many major brands are currently developing the technology, and some states within Australia have already begun conducting self-driving vehicles trials. Maintaining a focus on Australia, we find that over one thousand people are killed in traffic accidents each year, and more than thirty-five thousand suffer injuries which require hospitalisation. Moreover, traffic accidents impose serious pressure on public resources, costing Australia's economy, for instance, around thirty billion dollars per year. As approximately ninety percent of traffic accidents are attributed to human error, self-driving cars are poised to significantly reduce traffic accidents and impact the associated consequences, given that they are anticipated to outperform humans in many of the tasks involved in operating a vehicle. However, whilst self-driving cars promise so much, their development raises some serious ethical questions. I argue for a utilitarian approach to the problem of self-driving programming, focusing on the dilemma of unavoidable accidents in which all courses of action result in someone being harmed. In doing so, I make use experimental philosophy, evolutionary psychology, traffic accident data, and consider how the implementation of self-driving cars could have significant impact beyond traffic accidents. Given what is at stake, and the speed of technological progression, this is a serious ethical issue, and one that is ready to be addressed.