Researchers have published what is believed to be the first detailed examination of robot deception. They developed algorithms that allow a robot to determine whether it should deceive and designed techniques that help the robot select the best deceptive strategy to avoid getting caught.