Should your Tesla kill you?

As a sometime philosopher-theologian I come across some odd questions. Here’s one: ‘Should your Tesla kill you?’ – a recent twist on the classic ‘Trolley Problem’ thought experiment.

In case you’re blissfully ignorant or have happily forgotten Ethics 101, here’s how the Trolley Problem goes. A runaway train is heading towards five people stuck on the tracks. They will die unless you switch it onto another line with only one person on it. If you do nothing, five people will die – but not at your hand; if you flick the switch, only one person will die – and you’re responsible. What will you do?

Isn’t philosophy fun?

So, here’s the twist: if you’re in an accident, should your Tesla be programmed to protect the lives of you and your passengers, or innocent bystanders?

This is both more complicated and simpler than the trolley problem: in this case, I’m not killing someone to save others, but risking my own life to save them. Isn’t that what any morally responsible driver would do?

Not so fast. It matters that my Tesla’s not making an adrenaline-fuelled life-or-death decision, but enacting a kind of pre-programmed agency delegated to a machine and the algorithms that drive it. And frequently our cool-headed in-advance decisions are driven by self-interest instead – for example, the preponderance of over-sized 4WDs on suburban streets, which makes their drivers safer and everyone else less safe.

Whichever way we go, I wonder whether the true value of this thought experiment isn’t the answer we give, but how it exposes the ways such devices may distort our moral reasoning.

Outsourced protocols tend to muzzle our better angels in favour of individualistic self-interest. Sometimes our instincts and intuitions are better than our policies and protocols. And it’s hard to program them into a machine.

This Thinking Out Loud was first published on Facebook.