Our Robotic Future, Creeping Up

The other night I decided to watch the 2004 movie I, Robot. I had only seen it once before, in the theater 9 years ago. It popped up as I was browsing films in iTunes. While it's not the greatest SF or Will Smith movie I have ever seen, I enjoyed it, though I was skeptical of the level of AI and mechanical performance of the fictional year 2035 robots. Even before the latest super-smart NS-5 model that drives the movie's plot, the human-size, bipedal robots in wide use seemed to have good speech understanding and athletic ability, and even more impressively, they could interact seamlessly with humans in social situations ranging from home kitchens to bars to crowded streets. Could robot technology really reach that level 31 years from the film's release, just 22 years from now?

A few years ago, I might have said no, but now I'm not so sure. One data point: the recently announced Atlas robot from Boston Dynamics, which I read about today in this NY Times article. This 150 kg robot is hydraulically operated and is designed to work in human environments. It is an R&D model that will be programmed by several teams next year in a DARPA competition to demonstrate capabilities needed for human assistance missions. It's no cakewalk. From the Times article: "The contest involves programming the robot so that it is able to climb into a vehicle, drive to a destination, get out of the vehicle, cross a rubble field, open a door, use a power tool and turn a valve." Wow. You can watch the big lug in action in this DARPA video. It's clear this will NEVER be used for direct military purposes, despite the little hand exercise that looks like a wild west gunslinger spinning his 45's around. They probably only do that in movies anyway.

I'm not sure what to make of all this. I do believe that technology is developing at an exponential rate, and with sufficient computing power, strong but lightweight materials, and advanced batteries, I can imagine "Atlas" evolving into "Sonny" in 20 years. Sure. I've read a lot of other articles about how the use of robots is expanding into activities that require close interaction with humans, and how recent models can even be programmed by simply moving their arms to teach them the needed motions. Smarter, cheaper, easier to use robots are certainly not going to help our unemployment problems. Even if it never gets to the point of a robot bartender, a lot of jobs that could not easily be automated in the past will be up for grabs. Maybe even bartender.

Two things are clear. Boston Dynamics was thinking ahead when they loaded Atlas's chest with blue LED's. This is the same as the NS-5 in I, Robot, so when someone overrides their programming and makes them turn bad, it will be a simple matter to replace the blue lights with red bad-robot LED's (white hats and black hats are so twentieth century). The second thing is that when Skynet sends a Terminator from the future back to our time to locate some key technology, they should send it to Boston, and not Los Angeles.