This paper presents and evaluates a new distributed learning technology, called federated learning, and its applications in automotive systems. We review and classify existing approaches to federated learning, focusing on its implementation in connected vehicles. We also evaluate the challenges associated with this application domain. Federated learning allows data to remain within vehicles, thereby avoiding costly data transfers and mitigating privacy concerns. This technology shows promise in enhancing various automotive applications. Federated learning can be applied to driver assistance systems, predictive maintenance, and personalized user experiences in connected vehicles. By keeping data local, it supports privacy and reduces communication overhead. Federated learning represents a significant advancement in the use of connected vehicle data. It offers a practical solution to privacy and data transfer issues while enhancing the performance of automotive systems through collaborative model training.