Article seems a bit misleading. Apparently the robot is not self-aware (or proven to be) at all yet. They will just start testing to see if it will be soon. For all we know the robot may fail miserably.
Isn't google already self aware. Heck it knows what you are going to type before you type. Actually its aware of what you are going to think before you think.
also it's only image recognition. the robot will model itself and remember what it look likes, if the new image is different and the robot can recognize the difference it is self aware. yea um... not self aware imho.
What if the robot is self-aware -- self-aware enough to keep it a secret from us until he's ready to make his move?