The Autonomous and Five are hosting a Chapter Event on Safety & Artificial Intelligence. Through facilitated discussions among experts, we will work towards Global Safety Reference Solutions for safe autonomous mobility.
The event will explore the deep challenges associated with demonstrating – through verification, validation and safety argumentation – that an autonomous vehicle with AI-based components is acceptably safe for operation.
There is a fundamental conflict between traditional verification and validation, where a well-defined specification forms the basis of any safety argument and, for example, Level 4 autonomous vehicles where the very complexity inherent in the problem results in a lack of complete specification. Completely new methods of system architecture, implementation and verification are required.
During the event we will discuss two main questions: “How can we understand the competence of an AI component?” and “Can system-level mitigations protect against residual inadequacies in AI components?”
Tickets are still available! Interested? Please register below.