The tremendous technological developments in the automotive industry today are mainly fuelled by the development of vehicle-to-everything (V2X) communication capabilities and new automated driving features. Given their intrinsic requirements in terms of ultra-low latency and ultra-high reliable connectivity under high-mobility conditions, these features will only be unlocked over the long run with the large-scale adoption of 5G technologies. Among them, network slicing is considered as the key technology of an agile V2X use-case deployment, ensuring network flow isolation, resource assignment, and network scalability. However, while most deployments in Europe focus on evaluating the resulting network performance, security and privacy challenges associated with this technology have not been much investigated, notably in a cross-border scenario. Building on key 5G technologies (SDN, NFV) and machine learning algorithms (federated and deep learning), 5G-INSIGHT aims at: (a) proposing new techniques for road and network traffic prediction, thus allowing the early detection of intrusions and anomalies within 5G vehicular slices; (b) enforcing security-by design and privacy-preserving slicing policies for attack mitigation and personal data anonymization respectively; and (c) developing resource orchestration and management across multiple potential providers using federated slicing. The project will validate the proposed approaches by implementing simulations as well as a demonstration platform (Proof-of-concept) that will integrate the specific characteristics of the France-Luxembourg cross-border area.