The year is 2025. SpaceX has put a man and his dog on Mars. CO2 levels are reported stagnant. Things are good. You are weaving through traffic in your car to work and can just about eat breakfast, finish your presentation slides and practice in the next 10 minutes. It’s Friday; you’ll meet up with friends from uni later for a drink- or several. It’ll be one for the books. Surely you need to focus on the road, on driving? You don’t. Your car drives itself.
The Fault in Our (Driverless) Cars
Driverless or autonomous cars- which need little or no human control in operation- seemingly promise safer and smoother road interactions, but there are a number of ethical concerns with their operation. Firstly, in a scenario where an accident or even a fatality is inevitable, it is currently unclear whose safety the driverless car’s programming should be designed to prioritise- that of the passengers or the pedestrians. Furthermore, is the decision of priority to be made by the manufacturers or the car’s owner?
A classical utilitarian approach to tackling this concern demands that the option that benefits the greatest number of people ought to be taken. Marginal utilitarianism goes further and factors in the relative benefit to each individual. As such, an argument as to whether more potential victims outside the car compared to fewer passengers in it, becomes one on whether an old lady or a child should be saved. Nevertheless, as it is difficult to predict what cascade of events may occur due to an action taken by the system the utilitarian approach becomes nearly impossible to adhere to and equally for designers to create and implement in code. 76% of the respondents in an MIT survey on using driverless cars indicated that pedestrian safety should be prioritised over that of the passengers as this was the “moral” thing to do. However, they also indicated they would not buy or use a car that followed this mode of virtue ethics. This may explain Mercedes Benz’ decision to implement the “passenger-first” directive in its algorithm.
Just as concerning is the increased risk of hacking with growing Internet-of-Things (IoT) products and further digitisation of user information. In 2017 hackers were able to implement code into a fleet of BMW cars that caused them all to unlock simultaneously. They were then stolen. The company did not patch the error in its code till 6 months later. Too little, too late? In light of recent discoveries regarding automobile manufacturer fraud (see Volkswagen) perhaps the biggest worry consumers and legislators face is manufacturer ethics and transparency.
Another ethical issue that must be considered is the effect the use of driverless cars will have on the estimated 1.5 million professional drivers in the UK. If driverless cars are to become the standard, employers are more likely to use these safer, less likely to be tired and potentially cheaper alternatives which will in turn cause severely reduced wages and even complete job loss for those replaced. With possible unemployment looming in the horizon, is this a road we are ready to drive down especially with hands off the wheel?
Algorithm, take the wheel!
Autonomous vehicles, will almost inarguably limit yearly road accidents. 65% of all road accidents are estimated to be due at least in part to driver error. Between April and September 2017 alone, a reported 3430 people were killed in road accidents in the UK. The new driving system is controlled by integrated computer programming and sophisticated hardware designed to ensure cautious driving and strict adherence to traffic rules. Coupled with the fact that one of such vehicles is unlikely to have a cheeky pint at the pub or nod off on the way home, the driverless vehicles’ improvement of road safety is the most convincing argument in its favour. Companies like Google and Tesla are also leveraging data from test models in current use to enhance their responses to real-life situations. This also involves working on firewalls in the code that are impenetrable to outsider attack.
Technological improvements aside, the ethical argument may be tipped in the robot cars’ favour. Although utilitarianism suggests a compromise when appraising the degree of damage, it ignores the fact that the future is challenging to predict. Kant argues that if it is the right thing to do, it ought to be done. Few people would argue against saving the lives of over 1.2 million people worldwide who die in road traffic accidents yearly.
Whilst it is true that autonomous technology adoption will lead to the loss of conventional jobs; the transition will introduce new job opportunities in the emerging field. In the past we worried computers would replace workers in many industries, now most people work in businesses somewhat augmented by the use of computers. Should we start seeing autonomous cars as an opportunity and not a threat?
Proponents of driverless vehicles pose their own question: How many drivers on the road today were asked who they would act to protect in the event of a collision when getting a driver’s license? Three of the four writers of this article drive and confirm that this did not feature in their driving tests. An unreasonable degree of expectation is placed on engineers who design these vehicles and on the vehicles themselves. This is not to say that issues raised must not be considered, but critics- and supporters- must moderate their expectations.
The complexities in the adoption of autonomous cars have arguments from a number of traditional ethical standpoints previously mentioned. Modern technology is only growing and a post-modernist view may be considered: there is no consensus on individual ethical standards and automated vehicles indoctrination into society is inevitable. Germany has taken the lead in preparing a guide to the design and use of driverless cars and the world must follow. It will involve consideration not just from engineers but from social scientists and legislators alike to ensure that these potentially essential vehicles are as in line as possible with society’s ethics.