A Police Officer in Ottawa claimed to be in a "Robot-like-state" caused by post-traumatic stress disorder when he snatched a batch of groceries including 5 candy bars. This makes me wonder about the defense strategy that might be used by robot builders when their machine commits a crime. What will happen when the robot's programming is designed to use simple randomization or genetic algorithms to produce emergent, yet illegal behavior? If the courts can't be convinced that the programmer is completely responsible for the robot's behavior, will the robot be put in jail? Will the programmer be sent to jail even if their robot commits a crime that they did not specifically program it to commit? The possibilties of this future legal landscape are fascinating.