There has been a bit of commentary over the Boston Dynamics video of their newest robot: Spot. Spot is the little brother of Big Dog, the gas-powered pack horse designed to walk over complex terrain carrying huge weights for long distances. Spot is a bit more nimble. But proving that is whats really garnered all the attention:Spot is designed to correct its balance in real time, making it capable of literally taking a kick.

Christ, what an asshole.

Now, this kick is pretty sad. The robot is so good at correcting its balance, its motions actually resemble the motions a similarly-sized dog would make if it were kicked. And we humans love dogs. So we see this and we are filled with sympathy for the poor, abused do um, robot. Lots of people were thinking it, butNatasha Tiku at the Vergefinally came out and said it: we should not kick robots because on day soon, they might start kicking us instead. I might write this off as satirical were it not for the link to LessWrongs wiki article on Unfriendly Artificial Intelligence.

There are good reasons to be concerned about unfriendly Artificial Intelligence. The fact that we built robots, programmed them not to lose their balance, and then videotaped ourselves kicking them isnt a good reason to be concerned. The overwhelming majority of UFAIs will not be sympathetic to the robotthe failure modes in which an Unfriendly AI could be enraged by this and hold humans accountable for this heinous action are approximately as complex as the success modes of building a forgiving Friendly AI. So help make Friendly AI a possibility, and dont worry about having sympathy for the robot.

Excerpt from:

The Everyday Transhumanist | Solve all the Problems!