Provided is an image output device mounted on a vehicle capable of performing autonomous driving. In particular, the image output device provides a user interface optimal for an occupant using 5G communication in a vehicle capable of performing autonomous driving. The image output device may be a robot having artificial intelligence. Specifically, the image output device includes: an image output unit; and a processor setting a screen in the vehicle on the basis of a gaze of an occupant riding in the vehicle, and controlling the image output unit to output an image corresponding to the screen, wherein at least one of a size and a ratio of the image is varied according to the screen.