Recently I’ve seen scores of blogs and reports on the moral choices that will need to be programmed into the software that enables self-driving. These stories are a disservice and they are scaring the public into thinking that there is a chance that at some point a self-driving vehicle will make a programmed decision to kill them and they all neglect to cover at least two key points.
First, the opportunity costs of delaying deployment of self-driving technologies while we wrestle with developing the “perfect” response to these rare, imperfect, and artificially forced choice scenarios are more deaths, injuries, pain, suffering and property damage. Second, the level of situation awareness and decision-making that is and will be enabled by sensors, connectivity, information processing, and artificial intelligence far exceeds that of a human driver and will allow the vehicle to avoid very situations often caused by errors in human judgment that lead to the need make such moral choices.
These so-called “moral choice” questions are interesting cocktail chatter to be sure but let’s not lose sight of the fact that human drivers are far more likely to cause a fatal crash avoiding a deer or squirrel. And while the scenarios contemplated in these articles are interesting they are also very rare. Fatal crashes are not.
