Page - (000083) - in Autonomes Fahren - Technische, rechtliche und gesellschaftliche Aspekte

714.1 Why ethics matters Discriminating on the basis of age in our crash scenario would seem to be the same evil as discriminating on the basis of race, religion, gender, disability, national origin, and so on, even if we can invent reasons to prefer one such group over another. In Germany – home to many influential automotive companies that are working to develop self-driving technologies – the right to life and human dignity is basic and set forth in the first two articles of the very first chapter in the nation’s constitution [9]. So it is difficult to see how German law could even allow a company to create a product that is capable to making such a horrific and apparently illegal choice. The United States similarly strives to offer equal protection to all persons, such as stipulated in the fourteenth amendment of its constitution. If we cannot ethically choose a path forward, then what ought to be done? One solution is to refuse to make a swerve decision, allowing both victims to be struck; but this seems much worse than having only one victim die, even if we are prejudiced against her. Anyway, we can force a decision by modifying the scenario: assume that 10 or 100 other pedestrians would die, if the car continued forward; and swerving would again result in only a single death. Another solution could be to arbitrarily and unpredictably choose a path, without prejudice to either person [34]. But this too seems ethically troubling, in that we are choos- ing between lives without any deliberation at all – to leave it to chance, when there are potentially some reasons to prefer one over the other, as distasteful and uncomfortable as those reasons may be. This is a dilemma that is not easily solvable and therefore points to a need for ethics in developing autonomous cars. 4.1.1 Beyond crash-avoidance Many readers may object right away that the dilemma above (and others that follow) will never occur with autonomous cars. It may be suggested that future cars need not confront hard ethical choices, that simply stopping the car or handing control back to the human operator is the easy path around ethics. But I will contend here that braking and relinquish- ing control will not always be enough. Those solutions may be the best we have today, but if automated cars are to ever operate more broadly outside of limited highway environ- ments, they will need more response-options. Current research already makes this case as a matter of physics [12, 13], but we can also make a case from commonsense. Many ordinary scenarios exist today in which braking is not the best or safest move, whether by human or self-driving car. A wet road or a tailgater, for instance, may make it dangerous to slam the brakes, as opposed to some other action such as steering around the obstacle or simply through it, if it is a small object. Today, the most advanced self-driving cars cannot detect small objects such as squirrels [7]; therefore, they presumably cannot also detect squirrel-sized rocks, potholes, kittens, and other small but consequential hazards can cause equipment failure, such as tire blowouts or sensor errors, or deviations from a safe path.