The lack of acceptability of future autonomous vehicles (AVs) is a challenge for investing in their design. Driving simulators offer opportunities to imagine and test possible solutions to achieve acceptability. The solution studied here is a cognitive interface based on theories of mental representation, under the assumption that they are the basis of acceptability and trust. A first interface was designed and then tested by users on board a simulator. The interviews carried out at the end of the tests allowed characterization of user fears regarding self-driving vehicles and identification of ways of improving the initial concept. In particular, there is a need for information on the functioning and decisions of the AV. This information is essential for the construction of a mental representation that is sufficiently rich to establish a satisfactory level of confidence. However, information about AV functioning is not an end in itself, but a necessity to help in decision making processes according to the situation. For this reason, the second version of the interface will be connected to an empathic function in order to inform and reassure the user when his cognitive and emotional states require it.Read less <