Can computer programmers account for morality?

As Google's self-driving cars proliferate and the U.S. drone program grows, software engineers face the daunting task of anticipating — and coding for — moral quandaries

A cyclist rides past one of Google's self-driving cars outside the company's headquarters in California.
(Image credit: Justin Sullivan/Getty Images)

Within the next two or three decades, Google's self-driving cars will likely be commonplace, and our roads will be safer for it. Self-driving cars don't drink. They don't get tired after a long day at the office. And they'll have better reflexes, along with the ability to instantly communicate with one another to minimize accidents, leaving you free to text or tweet to your heart's content. Three states — California, Florida, and Nevada — have already deemed the vehicles legal, and it's only a matter of time before many of the other 47 follow suit.

But what happens to a self-driving car when it's time to make the morally unclear, split-second decisions that will inevitably come up on the road? It's a question Gary Marcus, a professor of psychology at NYU, tackles in a forward-looking New Yorker essay:

Subscribe to The Week

Escape your echo chamber. Get the facts behind the news, plus analysis from multiple perspectives.

SUBSCRIBE & SAVE
https://cdn.mos.cms.futurecdn.net/flexiimages/jacafc5zvs1692883516.jpg

Sign up for The Week's Free Newsletters

From our morning news briefing to a weekly Good News Newsletter, get the best of The Week delivered directly to your inbox.

From our morning news briefing to a weekly Good News Newsletter, get the best of The Week delivered directly to your inbox.

Sign up

That question will inevitably extend beyond our daily driving, too. Earlier this week, Human Rights Watch published a similar report tackling an innately more violent issue: Should military drones be able to decide who to kill? With the White House increasingly leaning on the drone program as a "counterinsurgency Air Force," one can reasonably assume that a growing percentage of the defense budget will be dedicated to drone R&D.

And the fact is, humans "aren't very good at codifying responses to dilemmas ourselves, particularly when the precise contours of a dilemma can't be predicted ahead of its occurrence," says Nicholas Carr at his blog. That ethical weight will fall heavily on programmers, whose code won't just have moral implications, but could be forced to face legal consequences as well. (Especially if your car decides to slam into the school bus after all.)

"We don't even really know what a conscience is," says Carr. "But somebody's going to have to program one nonetheless."

Chris Gayomali is the science and technology editor for TheWeek.com. Previously, he was a tech reporter at TIME. His work has also appeared in Men's Journal, Esquire, and The Atlantic, among other places. Follow him on Twitter and Facebook.