Ethics Of Self-Driving Cars

Algorithms of Life and Death: The Ethics of Self-Driving Cars

Research Question

Self-driving cars will inevitably have to face the decision of who to protect in the event of a road accident. Whilst human drivers are forced to make split-second decisions when confronted with an imminent accident, self-driving cars will be programmed with algorithms that will direct their actions when faced with such incidents. For example, when confronted with a wayward pedestrian, a self-driving car may have two courses of action: protect its occupant, or, take evasive action to save the pedestrian, though at considerable risk to the occupant's life. Thus, the core question of my proposed research arises:

How should self-driving cars be programmed to react

Indeed, major companies such as Google and BMW are currently developing such vehicles, with Google asserting that they aim to bring self-driving cars to the consumer market before 2020 (Tam 2012). Given that as much as ninety percent of car accidents are attributed to human error, a significant benefit of self-driving cars will be their great potential to reduce both serious injury and death on our roads (Kirkpatrick 2015). In Australia alone, we typically see over 1,000 people killed in traffic accidents each year, and over 35,000 serious accidents resulting in hospitalisation (BITRE 2017). With these amounts of injuries and deaths in mind, and the subsequent suffering that follows from such horrific events, the prospect of self-driving cars stands to be a significant occasion in terms of improving society and the lives of its members, not only in regards to direct harms, but also in terms of economic benefits that are likely to result from a decreased burden on public resources such as hospitals. However, although self-driving cars hold great potential for benefiting both individuals and society at large, their development raises serious ethical issues that need to be addressed before they are introduced onto our

Firstly, such research will be of great interest in the commercial sector, with manufacturers, if they are not already, having to face up to the tough question at the core of this proposal. Consumers will surely be concerned with the programming of vehicles which they, along with their loved ones, will travel within. That is, sales of self-driving cars that are set to sacrifice occupants whenever they are faced with an accident scenario in which the amount of individuals external to the car is greater than within, are less likely to become customers. Furthermore, the particular choice of algorithm adopted by a manufacturer may come under intense scrutiny in the event of a serious accident involving injury or death. Secondly, government will inevitably have involvement in the legislation surrounding self-driving cars. Given that accidents are unavoidable, and such accidents will be happening in the public sphere, government cannot avoid being a stakeholder in the ethics of self-driving cars. And finally, the general public, as previously alluded to, will want a clear understanding of what they are getting themselves into if they decide to purchase or travel within a self-driving car. Just as consumers want to know the water efficiency of the white goods they are purchasing,