Provided is a robot to which is disposed a cartilage vibration-transmission source, the robot being configured so that both hands of the robot are lain onto the face of a person or hold the back of the head, or so that one hand touches the cartilage of the ear. The relative positions of both hands touching the cartilage of both ears are maintained, but movement of the face is not restricted. The hands of the robot extend in the line-of-sight direction of the robot. The mouth of the robot is caused to move in conjunction with cartilage transmission audio. A limiter is applied to the touch pressure of the hands. Permission is sought before touching. Safety is confirmed before touching. When a malfunction occurs, touching with the hands is cancelled. The hands are warmed to the temperature of the human body.