The promise of a self-aware robot is almost a reality. Justin Hart and Brian Scassellati, PhD students from Yale University, are working on software to help robots recognize themselves in mirrors.
The robot is named Nico, and the pair are in the final stages of software recognition development. At the moment, Nico can only recognize its arm in a mirror, but Hart and Scassellati are optimistic that their robot will be ready for the final full “mirror test.” The final test is to mark Nico’s face with an odorless, non-tactile dye, place Nico in front of a mirror, and observe if it notices the dye.
The mirror test was developed in 1970 to see if certain animals were intelligent. To date, only a few non-human species have successfully passed this test, including some primates, elephants, and dolphins. Not even human babies can pass this test until they are 18 months old.
Hart and Scassellati plan to “teach” the robot to recognize its torso and head, along with shapes, colors, and textures. Hart said in an interview, “What excites me is that the robot has learned a model of itself, and is using it to interpret information from the mirror.”
In 2007, robots were developed to imitate movements and tasks in a mirror by classifying pixels of itself and other robots. Nico is different from these robots because it recognizes appearance and not motion, which is why having Nico pass the mirror test is so crucial to their research. Hart closed by saying, “This is an important step but it is not the endgame of artificial intelligence, it is just a step along the way.”
Hopefully Hart and Scassellati will also teach Nico about Isaac Asimov’s Three Laws of Robotics…