Bird's-eye view on the Automotive Cybersecurity Landscape & Challenges in adopting AI/ML

The integration of intelligent functionalities in con-nected and autonomous automotive system has great potential to deliver personalised user experience and improve traffic manage-ment. It can benefit the society by improving highway capacity and safety of road users. The adoption of data-driven Artificial Intelligence and Machine Learning models in the automotive sector is opening venues to new services and business models such as autonomous fleet management, self-driving trucks, robo-taxi etc. However, where the sharing of mix-critical data brings opportunities, it simultaneously presents serious cybersecurity and functional safety risks. In recent years, the cyber attacks have impacted every segment of automotive system including electronic control unit, infotainment, communications, firmware, mobile apps etc. This adoption of AI and ML as enabling technology for next-generation autonomous transportation systems is going to significantly widen the automotive attack surface. This trend has increasing tendency of exposing both vehicle and road -side infrastructure to a wide range of sophisticated cyber attacks. This paper aims to review and build a body of knowledge on the topic of automotive cybersecurity, by bridging a domain-specific knowledge gap among automotive system designers, engineers and system security architects. For this purpose, it discuss the autonomous driving system data processing pipeline and threat analysis and risk assessment process of automotive cybersecurity standard ISO/SAE 21434 to harness and harden automotive cybersecurity. It highlights automotive system architectural and ecosystem challenges in adopting AI and ML driven decision making.