AI: What’s a little car to do?!
It’s interesting to see the age-old problems of morality and ‘the right thing to do’ being applied in AI programming, outside the confines of hypothetical philosophical debates. As I tackle another HarvardX course on Justice and Morality, which begins with ‘the trolley dilemma’ to highlight differences between consequential and categorical moral reasoning – I can’t help but think re Machine Singularity. What do we teach an emotionless autonomous vehicle? What course of action is ‘right’? Should moral values be included in guidelines for machine behaviour? If so, should it be based on majority rule? How can we definitively answer these questions now, if the debates have been open ended since the dawn of time and Socrates!
A related idea is that of the value of a human life: whether a child has more value/right to live than the elderly? Is a life of a young mother more important than a single male? Or even more morbidly – how many elderly lives equal to one promising youth, if any?
Also at play are the responsibilities of the victims themselves: If two young mothers with their two children choose to cross a street on a red signal, breaking the law and getting into the car’s right of way, are their lives more or less important than the one innocent law abiding elderly man standing on the side walk? What about the person inside the vehicle? What’s an autonomous car to do?!
On March 18th 2018 the first self-driving car collision occurred when 49-year-old Elaine Herzberg crossed a shadowed road, at night, without a designated crossing, and was struck by a car on autopilot, whose driver was looking away. It’s a complicated situation, while the world is at the forefront of both regulating and designing these systems. The trolley problem is more relevant than ever before!
MIT’s MoralMachine game questions your moral roadside dilemmas:
Germany decides all lives are equal, and vehicles will not place any priority on age/sex, June 2017 vehicle ethics report: http://www.bmvi.de/SharedDocs/EN/publications/report-ethics-commission.html?nn=187598
Wired published a somewhat extensive guide re self driving cars, summarizing info and history on the subject:
Meanwhile Google’s Waymo automated taxi service is gearing up to launch:
For a philosophical perspective on the subject, read a study modeling ethics for self-driving cars, published in Frontiers in Behavioural Neuroscience: