Seite - (000089) - in Autonomes Fahren - Technische, rechtliche und gesellschaftliche Aspekte

774.2 Scenarios that implicate ethics tialist decision for the robot car – if all we care about is maximizing lives and minimizing deaths – is apparently to drive off the cliff and sacrifice the driver, since it is better that only one person should die rather than more than one, especially 10 or all 30 persons. This decision would likely be different if, instead of a school bus, your robot car were about to collide with another passenger car carrying only one person. Given the same average odds of death, one in three, the expected number of deaths in a collision would only be 0.67, while the expected number of deaths in driving off a cliff remains at one. In that case, the right consequentialist decision would be to allow the accident to occur, as long as the average odds of death are less than one in two. If, instead of another vehicle, your car were about to collide with a deer, then the decision to stay on the road, despite an ensuing accident, would be even more obvious insofar as we value a deer’s life less than a human life. Back to the school-bus scenario, programming an autonomous car with a consequential- ist framework for ethics would seem to imply your sacrifice. But what is most striking about this case might not even be your death or the moral mathematics: if you were in a manually driven car today, driving off the cliff might still be the most ethical choice you could make, so perhaps you would choose certain death anyway, had you the time to consider the op- tions. However, it is one thing for you to willingly make that decision of sacrifice yourself, and quite another matter for a machine to make that decision without your consent or foreknowledge that self-sacrifice was even a possibility. That is, there is an astonishing lack of transparency and therefore consent in such a grave decision, one of the most important that can be made about one’s life – perhaps noble if voluntary, but criminal if not. Thus, reasonable ethical principles – e.g., aiming to save the greatest number of lives – can be stressed in the context of autonomous driving. An operator of an autonomous vehicle, rightly or not, may very well value his own life over that of everyone else’s, even that of 29 others; or he may even explicitly reject consequentialism. Even if consequential- ism is the best ethical theory and the car’s moral calculations are correct, the problem may not be with the ethics but with a lack of discussion about ethics. Industry, therefore, may do well to have such a discussion and set expectations with the public. Users – and news headlines – may likely be more forgiving if it is explained in advance that self-sacrifice may be a justified feature, not a bug. 4.2.3 Ducking harm Other ethical principles can create dilemmas, too. It is generally uncontroversial that, if you can easily avoid harm to yourself, then you should do it. Indeed, it may be morally required that you save yourself when possible, if your life is intrinsically valuable or worth protect- ing; and it is at least extrinsically valuable if you had a dependent family. Auto manufac- turers or OEMs seem to take this principle for granted as well: if an autonomous car can easily avoid a crash, e. g., by braking or swerving, then it should. No ethical problem here – or is there?