Causal Liability in Autonomous Systems

With the widespread use of autonomous systems, liability of-ten shifts towards the manufacturers and part suppliers, especially whenvarious system components, sourced from different suppliers, contributeto a failure. This necessitates a framework for automatic liability apportionment which can be embedded in manufacturer-supplier contracts andminimise legal disputes. To this end, we propose a formal framework basedon the notion of actual causality in structural causal models and the robustness semantics of logical specifications. We prove several desirableproperties of this framework. Moreover, we formalise the notion of causalnon-interaction, give sufficient conditions for it to hold, and demonstrateits utility in deriving upper bounds for liability analysis. Furthermore, werelate our definition to the existing notion of harm, empirically evaluateour framework to demonstrate its efficacy, and release a software packageimplementing our approach. We extend our framework to handle intervalspecifications.