Scientists at the Rensselaer AI and Reasoning Lab (RAIR) in New York experimented with three NAO robots to demonstrate self-awareness.
The three robots were programmed to believe that they were given 'dumbing pills' and could not talk. In reality, two were muted and the third was not. When asked if they were given the pill, one robot spoke up, "I don’t Know." Then, he recognized his own voice. The scientists theorize that this proves the robot to be self aware.
Read more, here.