Connected and self-driving cars have emerged over the last decade as a leading example of cyber-physical systems, which seek to considerably enhance traffic safety, reduce emissions, decrease costs, and improve efficiency. Google, TESLA, Uber ATG are becoming pioneers in the autonomous vehicles industry. Autonomous vehicles can have a large codebase and with a large volume of messages exchanged. The concepts of connected driving, cooperative driving, and intelligent transportation systems increase the connectivity of vehicles to the internet or other cloud services. High connectivity, misconfiguration, and insecure coding widen the attack surface of autonomous vehicles. A conceptual model of autonomous vehicles is presented in this paper to better understand autonomous vehicles' attack surface. In addition, specific threat modeling techniques are discussed. Experiments were carried out to demonstrate the security risks to autonomous vehicles due to third party Electronic Control Units.