If you’ve watched the various Boston Dynamics videos released over the last few years you’ve already seen the 4 legged robot called “Big Dog” that can keep its balance when being kicked from one side. Then there was the fast-running variation called “cheetah” which could run at up to 28 mph. Both of those were impressive but they weren’t humanoid and didn’t even have anything that looked like a head. That made it easier to see them as impressive machines.

Atlas is different. He not only has a head but hands that can pick things up, though he lacks discernible fingers yet. What he can do doesn’t seem dramatically more advanced than previous robots (though getting up off the ground is an impressive trick I haven’t seen before) but there’s something about seeing this 6 foot tall humanoid figure with a spinning head that feels different. I listened to one newscaster (CNN?) saying she felt a little sorry for Atlas after watching him get pushed around in this clip. I had a different reaction. Watching that engineer knock a box out of its rubber coated hands and then shove him backwards with a hockey stick has me wondering how long it will be until Atlas can respond by smacking the stick out of the guy’s hands and shoving him back a lot harder.

Boston Dynamics was initially funded by DARPA and the military applications were part of the early designs. Now that the company is owned by Google X (a subsidiary of Alphabet) it’s not clear if this is being retooled as a civilian robot (again that’s what the video seems to suggest) or possibly some kind of rescue robot. But how long will that last?

Even if Google refuses to “be evil” by turning Atlas or his progeny into terminators, won’t someone else come along who is more than willing to scoop up those military contracts and mount an M-4 on this thing? Consider that the current debate over how to deal with ISIS in the Middle East and Africa revolves around the desire to avoid putting American “boots on the ground.” The idea is that, yes, this is an important threat we need to deal with but not important enough to risk American lives over. How does that dynamic change once it becomes possible to send in an army to fight ISIS that doesn’t need boots and doesn’t risk any American lives?

The logic of it seems inevitable. We already have a president with a kill list and an air force full of remotely piloted drones. What’s so different about having a ground force of drones with legs? One could even argue that sending a team of robots to take out an ISIS leader is more humane than firing a missile at an unknown number of people on the ground. At least with Atlas there’s less potential for collateral damage. Some military “pilot” sitting in safety hundreds of miles away could walk in and see exactly who they are killing and avoid shooting anyone who isn’t on the President’s list.

Just a few years ago all of this seemed, to me at least, like a plot for a summer action movie, not something likely to happen in the near future. Watching this video, it seems a lot closer now. We’re not there yet, but in another 10 years we really could be having a national debate about the role of robots in combat.

[Credit where due: The headline is a joke I’ve seen Ace of Spades use a hundred times. I’m borrowing it for this post.]