With recent decades seeing technological advancement occur at unprecedented rates, it should come as no surprise that we constantly strive to use new tech to further improve our quality of life as a species.\u00a0One cutting-edge application of modern technology lies in the realm of artificial intelligence (AI), which possesses the power to automate tasks of almost every difficulty level. Given its the current trajectory of development, \u00a0AI could feasibly be used to automate transportation – a field which has been the subject of consistent advancement in recent years. Current statistics report an average of 9.1 crashes per million miles driven in autonomous cars, versus 4.1 for human drivers. However, many accidents involving self-driving cars can be attributed to unpredictability on the part of other human drivers, suggesting that an entirely autonomous world of transportation – and the associated intercommunication between self-driving vehicles – could bring accident rates far below those of most human drivers. Furthermore, accidents involving AI-driven cars tend to be less destructive than traditional collisions, with fewer injuries sustained by those involved. It seems almost inevitable that AI technology will soon revolutionize transportation as well know it, from passenger to commercial services.<\/p>\n
As we discussed in class, one possible objection to the large-scale introduction of autonomous vehicles lies in the realm of morality. Does AI possess the ability to properly evaluate high-risk situations with the same moral considerations as a human brain? Are we humans better able to weigh the morality of, for example, hitting a child as opposed to an elderly citizen? Can we better tailor our behavior to reflect known human values inaccessible to AI?<\/p>\n
I personally find the above objections entirely unconvincing. Before even considering the ethics of collision behavior, one must recognize that an interconnected network of AI-driven vehicles would be uniquely able to govern everyday driving decisions – such as who should proceed first at an intersection – in a manner which best minimizes crash potential. Humans are imperfect, and there is simply no way for individuals in separate vehicles to communicate their intentions as effectively as computers. If a crash does occur, however, why does the notion of AI deciding who might live and who might die inspire such panic among human moralists? I would propose that the simple lack of control over an outcome scares people most. And yet, given a split-second life-or-death decision, panicked humans could not possibly reproduce consistent outcomes. To use an aforementioned example, some would instinctively swerve to hit the child, some would hit the elder, and many would be completely incapable of reasoning through the issue. This is because there is no universal “right” answer to the question of how tragedy should best be averted; how can we expect AI to make the correct decision in every high-risk situation when we ourselves cannot identify what constitutes “correct”?<\/p>\n
And this does not even begin to touch on the matter of passenger safety. I would personally be the first in line to purchase a car which puts the safety of my children or passenger before my own in a crash. In the specific case of AI-driven vehicles, and also in general,\u00a0I would much rather devote consideration to the technological challenges of AI use – the potential of self-driving cars to be hacked, for example – than to worry about a “morality” which we humans cannot ourselves define.<\/p>\n
Additional sources:<\/p>\n
15 Unbelievable Self-Driving Car Statistics & Facts (2019)<\/a><\/p><\/blockquote>\n