GenAI can become a game changer for the automotive industry. One of the most promising application areas is the generation of code. Stunning results have been observed and various teams are striving to apply GenAI in development processes. The potential for efficiency improvement is huge. How does this fit into Automotive development where software is developed based on SW requirements, SW architecture and SW detailed design? It is currently not clear how information about requirements or software architecture can be systematically brought into the code generation process using prompts. There are several ways to handle this: Waive the Automotive development philosophy completely and ignore requirements and architecture? Or better adapt Requirements Engineering and Architecture Design to fit to GenAI? An analysis reveals that there is also the promising opportunity to train a specific AUTO-AI that only knows one proven Standard Software Architecture. In this presentation we will investigate the options and evaluate their potential benefits. Conclusions: AI is here to stay. However, the integration into proven development systems needs serious consideration. If we stay with our proven development philosophy of system engineering, we must find a way to bring the intended software architectural design into the GenAI code generation process. An AUTO-AI should be developed by a consortium of partners to generate code according to a Standard Automotive Architecture.