The further advancement of autonomous driving, which strongly relies on machine learning based perception systems, requires rethinking the communication between the vehicle and the user. Combining information from the environment perception with in-cabin monitoring and especially the perception of the driver’s reactions brings together important aspects in autonomous vehicles. Departing from different perception models and comparing models for detecting bodies outside the vehicle with far more detailed models for in-cabin detection, the presentation will discuss how external perception systems communicate with the user and what improvements should be achieved in the future in autonomous vehicles.