Ethics in Autonomous Driving

...

case, to my mind, robotic cars can still be defended against unethical charges. ~35,000 lives were arbitrarily and for no obvious reason condemned to be victims this year, thus there is no moral issue with replacing them with a new subset of unfortunate future victims. Only if there was a pre-identifiable set of people, whose risk of falling victim would be higher only due to the introduction of self-driving cars, Lin’s “non-identity problem”-related ethical concerns would be justifiable.

3.2 Utilitarian/Consequentialist view

Supporters of utilitarianism, an altruistic form of consequentialist view defines that the “right action must maximize overall good”. The theory, based on philosophers Jeremy Bentham and Stuart Mills work, points out that being ethical means to maximize good or minimize bad for the entire human community. Therefore “the purpose of all moral action is to achieve overall happiness”.9

Hence follows, a consequentialist would support autonomous cars unconditionally as any net savings of lives leads to a positive result from an ethical point of view. As mentioned before, experts agree that overall, there is a very positive outlook on traffic-related death tolls through the introduction of robotic cars. Thus, utilitarian bigger picture thinking points out the maximization of the overall good of decreased mortality rates.

Consequentialists try to reduce the variety of relevant ethical considerations like duties or rights of individuals to only one single principle – maximizing general utility. The problem with that theory becomes obvious when studying philosopher Foot and Thomson’s “Trolley Problem” , which is absolutely applicable to autonomous cars. The “Trolley Problem” debates if a train should be shunt on a side track and kill one person, instead of staying on the rail where five people are standing. The act utilitarian would certainly agree in order to reduce overall harm and thus a robot car should be programmed to always follow this guideline in case of an unavoidable crash. The downside here is that there is a significant difference between “letting die” five people and actively “killing” even just one person.11 Programmers of robotic cars would have to resonate with that in order to secure maximized utility at all cost. Therefore consequentialism here does not provide sufficient ethical guidance.

3.3 Kantian/Deontological view

Deontological ethics adapted from Immanuel Kant put “duty” at the center of consideration. Only if reason, instead of inclination guide someone’s will, correct moral principles can be derived. “The moral worth lies in the act itself and not in (…) happiness or the consequences”. Consequently, universal principles should govern our actions and therefore Kant formulated the categorical imperative – which can be loosely translated in: “act such that the principle of one’s act could become a universal law of human action (…).”

Hence follows, the Kantian view on autonomous driving is clearly objecting the consequentialist view. Of course, accidents and fatalities through bad luck are justifiable in the eyes of Kant, yet programmers who define algorithms in a way that intends to keeping the death toll low at all cost are morally incorrect. Hevelke and Rümelin (2014) emphasize that “a violation of some person’s fundamental rights cannot be legitimized on the basis of benefits for others, no matter how large.” Deontological ethics object that any innocent person would be sacrificed as a means to an end, even if that end is saving five people.

Again, also Kant fails to give clear guidance for automakers on how to program robotic cars. In those complex situations, with possible knock-on effects, “two principles may conflict without there being a clear way to decide which (…) rule should take precedence.”

4.) Personal recommendation for robotic car programming

First of all, I would like to underline that I absolutely support the development of autonomous vehicles and the technological progress will benefit our lives. Automated cars promise great benefits and a change in road transportation will be inescapable. Anuj K. Pradhan, research scientist in the Human Factors Group agrees that ethical studies are crucial and timely, but they ought to be balanced with real-world applications. "I do not think concerns about very rare ethical issues of this sort [...] should paralyze the really groundbreaking leaps that we are making in this particular domain in terms of technology, policy and conversations in liability, insurance and legal sectors, and consumer acceptance," Pradhan pointed out. The focus of research must clearly be on technological advancement that will in the end lead to the expected benefits like a reduced death toll. Human error and carelessness is still the number one reason for fatal car accidents and also human drivers often face tricky situations in which moral considerations are needed. However, “human drivers who face these situations may not even be aware that they are [facing a moral situation], and cannot make a reasoned decision in a split-second”, and therefore those considerations are considered new and unique, French psychological scientist John Bonnefon realizes.

Nevertheless, besides all focus on machine learning improvement, the automotive industry must inevitably answer one key question: should autonomous cars ever sacrifice their owner for a potential greater good? The problematic nature of this question becomes explicit when reflecting on the study published at MIT recently: a vast majority in six Amazon Mechanical Turk studies approved of utilitarian autonomous vehicles that sacrifice their passengers for a greater good, “and would like others to buy them, but they themselves prefer to ride in AVs [autonomous vehicles] that protect their passengers at all costs.”

The study proved to me, that robotic cars can only succeed if they protect their users at all costs. Interestingly enough, only a few days ago, German car manufacturer Mercedes was the first industry player who made this confession very clear. Executive Christoph von Hugo pointed out, “if you know you can save at least one person, at least save that one. Save the one in the car. If all you know for sure is that one death can be prevented, then that’s your first priority.” According to von Hugo, this proposition is non-negotiable, as the consequence of a utilitarian sacrificing of the car is often too complex to predict. Swerving the car in order to avoid a group of children might lead the car to crash frontally into an approaching school bus. Thus, Mercedes will best protect its cars by working hard on avoidance of crash situations. “We

...