Autonomous vehicles are an emerging technology that are expected to reduce traffic and congestion, while increasing road safety. However, an ethical dilemma exists when coding the vehicles behaviour. In the event of an unavoidable crash that would kill either pedestrian(s) or the passenger(s) in the car, how should the car be coded to respond? Either choice has implications that are likely to affect both public health and public opinion on autonomous vehicles.
Prioritisation of occupants
Although controversial, saving the cars occupants in all situations would potentially result in greater long term gain. Autonomous cars could eliminate 90% of road traffic fatalities if they replace conventional cars. Using 2013 statistics, there would be 1541 fewer direct fatalities per year in the UK alone and over 1.1 million fewer deaths per year worldwide, greatly benefiting public health and society.
Despite general support for utilitarianism ethics (doing the least amount of harm in total) in the programming of autonomous cars, the public consensus is that personally they would not purchase a car which was programmed in some situations to sacrifice the occupants. This would reduce the usage of autonomous cars, increasing the number of deaths from normal road traffic accidents costing more lives.
Additionally, a greater uptake of autonomous cars would result in a reduction in congestion, reducing time spent in traffic, increasing productivity, reducing air pollution (which would reduce the 40,000 premature deaths a year linked to this issue and help to provide independence and mobility to disabled people. If the uptake of autonomous cars is slower, these benefits will be delayed. To promote autonomous vehicle uptake, in the extremely rare case of an unavoidable fatal crashes, priority should be given to the vehicle occupants.
Recent empirical studies suggest that many laypeople would not object to a program that makes autonomous vehicles swerve and kill a bystander in order to save more people from a potentially lethal accident. If the car contains a number of occupants (potentially including children) it may be more appropriate for the car to sacrifice a bystander in this situation.
For a driver facing the situation of an unavoidable crash, it is likely that the course of action the driver would take would favour self preservation due to natural survival instincts. As humans have a reaction time, they would not have time to weigh up options or outcomes – so the natural reactions would take over. Coding the car to always kill the car’s occupants and save pedestrians would rob the driver of their ability, and choice, to attempt to save themselves. In a statement, Iyad Rahwan from the Massachusetts institute of technology summarises this by saying “people want to live in a world where cars minimise casualties, but everyone wants their own car to protect them at all costs”
Prioritisation of other road users
Road safety is always the first and foremost obligation that a driver should consider. Protection of other road users is essential to ensure that the general public are living in a safe and harmonious environment. This also emulated within ethical theories. An example of which is utilitarianism which states that the principle of utility is the ultimate principle of morality. Practically, this means choosing whichever action or social policy would have the best consequence for everyone concerned. In the case of an unavoidable accident, the concern for the majority of people is the existential risk to human life. In highly populated areas such as cities, the number of pedestrians will be larger than the number of cars, hence saving pedestrians protects the majority of people which agrees with the utilitarian approach.
This idea is also supported by Kantian ethics, which focuses on the responsibility rather than the resulting outcome of an event. Despite the increasing uptake of autonomous car usage, which is resultant from prioritising the cars occupants, there should be no reason to increase the danger to pedestrians and fellow road users. Following Kantian ethics, it is the drivers responsibility for the actions of their car in the event of an unavoidable accident and thus the pedestrian should be prioritised.
As described by a survey in The Guardian, people are less willing to buy an autonomous car if it is programed with a moral algorithm which would sacrifice its occupants rather than others. Some people are worried this algorithm would become an obstacle toward the popularisation of autonomous cars, which may induce more accidents. This argument is not persuasive because the driver can be a pedestrian in other circumstances. Although it seems like the moral algorithm is threatening the driver’s own life, it has made the most appropriate decision to protect everyone’s interest. In other words, the driver may possibly be saved by the moral algorithm while he/she is the one on the street when accident occurred. The algorithm is fair to everyone as its always choose the action that minimize the loss of lives so there is no reason to avoid using an autonomous car. To sum up, saving the pedestrians does not mean threatening the driver’s life.
No matter who is either inside the autonomous car or the pedestrian, the value of life of each person should be the same based on equality. Therefore, killing less people should be the best choice for our society to minimize the total amount of harm. This is a similar outcome to the trolley problem thought experiment, where sacrificing the life of one person saves the life of many others.
In conclusion, the benefits of autonomous car uptake and the potential socio-economic gains must be balanced against the life of all road users and the responsibility of autonomous car users in placing others at risk.