Driving Trust: Ensuring Automotive Safety in the Age of Artificial Intelligence

Counts differ because each database indexes a different set of publications. We treat OpenAlex as the canonical count; Google Scholar is not shown (no API, and crawling it violates its ToS).

“Superintelligent machines may soon surpass human capabilities, posing unprecedented threats if safety is not prioritized over corporate competition.” — Dr. Geoffrey Hinton This quote from AI godfather Dr. Geoffrey Hinton perfectly reflects the growing concern around the rapid rise of Artificial Intelligence. Over the last few years, AI has evolved from being just a research topic into a technology that is now becoming part of our everyday lives. From smartphones and healthcare to automation and modern vehicles, AI is slowly influencing the way systems make decisions and interact with humans. This manuscript presents a comprehensive and forward-looking perspective on the evolving field of AI safety in the automotive domain. It discusses both the advantages and challenges of using AI in automotive systems, with a particular focus on safety. The manuscript highlights AI-related incidents through a literature review and explains how incorrect or unexpected perceptions by AI-based systems can lead to unwanted hazards and potentially threaten human life. It also provides a futuristic perspective on the race among industries to adopt and advance AI technologies, and how this race could potentially lead to unsafe scenarios, using the Swiss Cheese Model to understand how multiple contributing factors can result in an unsafe event. The manuscript further provides an overview of how AI-based systems differ from traditional systems, discusses the key challenges associated with these differences, and explores possible ways to address them. It also briefly introduces emerging standards in the field of automotive AI safety, particularly ISO/PAS 8800, and discusses how such standards can help improve the safety of AI-based automotive systems. Keywords: AI safety, Artificial intelligence, Machine Learning, Deep Learning, ISO/PAS 8800, ISO 26262, ISO 21448, AI incidents