Remember how last month Qbo learned to recognize itself in a mirror? The big question after that little feat was what would happen if Qbo met another Qbo: would it think it was just looking at itself in a mirror, or would it recognize another model as a unique individual? Well, now we know.
I'd just like to point out that two of our commenters on the Qbo mirror video, Robotbling and Bevan, speculated that Qbo would use exactly this method to differentiate itself in a mirror from another, similar robot: By taking a series of actions (in this case, flashes with a nose LED) and watching whether the other robot responds identically, Qbo is able to determine whether it's looking at a reflection or not. Smart, guys!
As cool as this demo is, it doesn't really go all that much farther towards settling the question of whether or not a robot can be self-aware. From one perspective, Qbo now has more of a sense of individuality than it did before, but only because it has been programmed that way by a clever human. Personally, I'm of the opinion that this is all fun to think about, but robots are still just running code that we've instructed them how to run. I'm tempted to say that assigning consciousness or self-awareness to a robot would require some sort of emergent behavior that couldn't be traced back to a sensor input or algorithm, but on the other hand, arguably human behaviors are simply the result of algorithms processing sensory information. I guess at the end of the day we may just have to fall back to the ol' Turing Test: If a robot acts self-aware in such a way that we can't distinguish it from a human acting self-aware, then there we go, we've created at least a limited form of artificial consciousness.
[ TheCorpora Blog ]
Evan Ackerman is a senior editor at IEEE Spectrum. Since 2007, he has written over 6,000 articles on robotics and technology. He has a degree in Martian geology and is excellent at playing bagpipes.