Ethical Killing: Programming Autonomous Vehicles
29 Nov 2017The Autonomouse Trolley
It is a ethical thought experiment that nearly everyone heard before. A trolley is out of control headed down to a fork in the tracks; you are in charge of the controls. If you do nothing it will continue down its path where you see several people, who if hit, would surely perish. If you choose to change the course a single person you see on that side track will surely perish. What do you do?
This question has become more pertinent as the development of autonomous vehicles is exploding, and accidents are inevitable. Unfortunately the problem arises that at some points we may be forced to violate our obligation as engineers to “avoid harm to others” (ACM Code of Ethics), but is there an ethical way to do that? One of the most prevalent arguments seems to be to “act consistently with the public interest” (IEEE Code of Ethics) which is a train of ethical thought most synonymous with utilitarianism. Utilitarianism is the ethical approach with the idea of doing the most good, or least harm.
An interesting obligation to play out in our quest for autonomous vehicles is the obligation to “Ensure that users and those who will be affected by a system have their needs clearly articulated during the assessment and design of requirements; later the system must be validated to meet requirements”(ACM Code of Ethics). We can look at the people hit by the trolley, in our case people affected by our system. Did they have a say in the assessment or design? Probably not, yet it affected them severely. How then can we justify this as we would have a moral and legal obligation to do so.
As hard as these questions are to answer, I don’t believe they are meant to be answered by engineers. It will be answered in the courts, by the public and the courts. Which I think is ethically right as then the public will determine for engineers what is good for them. Thereby allowing engineers to fulfill the needs of the many, such as utilitarianism dictates. Regardless of these specific ethical cases autonomous cars are here and their established track record for safety is already an improvement upon the status quo, thus doing the most good.