The Ethics of Driverless Cars

It’s starting to feel like the trajectory towards autonomous vehicles is becoming less possibility, more inevitability. 

It’s not just that Google and others with deep pockets are trying to crack this nut, but so are governments. The latter range from simply accepting the inevitability of sharing our roads with HAL 9000 all the way up to attempting to position their jurisdictions as a leading innovator. I've seldom seen legislative bodies so far ahead of the curve trying to anticipate an industry that doesn't exist yet.

An interesting narrative has come out of all this: what are the ethics of driverless cars? What's being widely discussed is how to programatically replace humanity's capability of making snap moral judgements. The thinking goes something like this:

• HAL knows a crash is inevitable. There's a group of children who have jumped in front of the car at a distance less than its maximum reduction in velocity per second (that's slamming the brakes to you luddites who still drive with a steering wheel). The only evasive manoeuvre is swerving, which would make roadkill of an 95-year-old with terminal cancer. What should HAL do? Or specifically, how should the guy who programs HAL approach this?

This is classic philosophy: straight up Jeremy Bentham versus Emmanuel Kant. And, also not the subject of this post.

The two dozen pieces I've read around the ethics all gravitate to variations of the above problem, and to an extent who pays when these problems wind up in a lawsuit.

Still, I feel there are some equally interesting problems which (I've not seen) enter the debate, which I've listed below. Some of these assume not just autonomous vehicles sharing the roads with us, but assume that the full road network has become a fully-automated ecosystem. Go watch any sci-fi movie, ever, of your choice if you don't get what I mean by that.

• Apple vs. FBI. No, not the iPhone. The FBI agents chasing the bad guy with a top-of-the-line iCar in a high-speed automated pursuit. Does law enforcement get to pull over HAL at will if they want? Should HAL auto shut-dow as a programmatic response to sirens (or the BlueTooth version of a Siren)?
• Bending the Rules of the Road. I've read stories about a car being pulled for speeding, to turn into a full police escort - for the lady on her way to the hospital to deliver. Or what about "driving with the flow of traffic", which in Southern California means Speed Limit +10. Should HAL make a judgement call here?
• Internet Road Neutrality. In a full-network of autonomous cars, should all cars be treated equally, by the same rules, or should road providers (like ISPs in today's debate) be able to charge for premium speeds and service? Is travelling 100 km/h a basic human right?
• Animal versus Car. Should HAL swerve to avoid a cute, furry animal at no risk to the car to avoid killing it, even if the car is not at risk by doing so? Probably. Should it do the same for a mosquito or a Poodle about to hit the windshield? Clearly not. Where is the line drawn? What animals will Noah the programmer of HAL favor?
• Overclocking & Modding. Back in the day, people used to own the hardware they bought, and could tinker with it as they saw fit. Should you be able to mod your car so, maybe, it favors saving you instead of saving others? As in, the car can choose between swerving off a cliff, or hitting a group of 100 nuns on their way to be canonised - and you have a personal preference for the latter?
• Banning Humans. If we can categorically prove (and I bet we will, soon) that driverless cars cause 90% less deaths and injuries from car accidents then humans do, should we ban humans from driving? That'd make the difference between HAL and a sober human far bigger than the difference between a sober human and a drunk one.

Have fun with that list. I'm no ethicist, but sure am excited to see how these turn out in the end.