Seite - (000087) - in Autonomes Fahren - Technische, rechtliche und gesellschaftliche Aspekte

754.2 Scenarios that implicate ethics matters what kind of vehicle we would crash into (e. g., is it a compact car or SUV?), how heavy incoming traffic is (e. g., would more than one vehicle be involved?), how many persons may be involved (e. g., are there children in the other car?). Of course, here we are assuming perfect sensing and V2X communications that can help answer these questions. If we cannot answer the questions, then we face a possibly large unknown risk, which makes driving into incoming traffic perhaps the worst option available. Other factors relevant to the decision-points above include: the road-shoulder type (paved, gravel, none, etc.), the condition of the car’s tires and brakes, whether the car’s occupants are seat-belted, whether the car is transporting dangerous cargo that could spill or explode, proximity to hospital or emergency rescue, damage to property such as houses and buildings, and more. These variables influence the probability of an accident as well as expected harm, both of which are needed in selecting the best course of action. From this short analysis of a typical crash (or possible crash) with an animal, we can already see a daunting number of factors to account for. Sensing technologies today can- not answer some or many of the questions above, but it is already unclear that braking should be the safest default option – as a proxy for the most ethical option – given these uncertain conditions, all things considered. Automated cars today can already detect whether there is oncoming traffic in the opposite lane. Therefore, it is at least possible that they can be programmed to maneuver slightly into the incoming lane under some conditions, e. g., when there are no incoming cars and when it may be dangerous to slam on the brakes. Whether or not sensing technologies will improve enough to deliver answers to our questions above, a programmer or OEM would still need to assign costs or weights to various actions and objects as best as they can. Yet these values are not intrinsic to or dis- coverable by science or engineering. Values are something that we humans must stipulate and ideally agree upon. In constructing algorithms to control an autonomous car, ethics is already implied in the design process. Any decision that involves a tradeoff such as to strike object x instead of object y requires a value-judgment about the wisdom of the tradeoff, that is, the relative weights of x and y. And the design process can be made better by rec- ognizing the ethical implications and by engaging the broader community to ensure that those values are represented correctly or at least transparently. Working in a moral bubble is less likely to deliver results that are acceptable to society. Again, in a real-world accident today, a human driver usually has neither the time nor the information needed to make the most ethical or least harmful decisions. A person who is startled by a small animal on an otherwise uneventful drive may very well react poorly. He might drive into oncoming traffic and kill a family, or oversteer into a ditch and to his own death. Neither of these results, however, is likely to lead to criminal prosecution by themselves, since there was no forethought, malice, negligence, or bad intent in making a forced, split-second reaction. But the programmer and OEM do not operate under the sanc- tuary of reasonable instincts; they make potentially life-and-death decisions under no truly urgent time-constraint and therefore incur the responsibility of making better decisions than human drivers reacting reflexively in surprise situations.