A common belief regarding current safety engineering practices is that they are inadequate to address the complexities of AVs and the data-driven nature of AI models. Therefore, some argue, a new approach to safety engineering practices is needed to rigorously assess and assure the behaviour of AI components when traditional fault-based approaches fall short, but these new approaches present new challenges in untested waters. This session will highlight the nature of how those conflicting views on AV safety develop and offer both practical solutions and technical strategies for companies to complete safety development automated driving, with an overview of ISO 8800, ISO 5083, ISO 21448, and ISO 26262.