Changes and updates in the requirement artifacts, which can be frequent in the automotive domain, are a challenge for SafetyOps. Large Language Models (LLMs), with their impressive natural language understanding and generating capabilities, can play a key role in automatically refining and decomposing requirements after each update. In this study, we propose a prototype of a pipeline of prompts and LLMs that receives an item definition and outputs solutions in the form of safety requirements. This pipeline also performs a review of the requirement dataset and identifies redundant or contradictory requirements. We first identified the necessary characteristics for performing HARA and then defined tests to assess an LLM's capability in meeting these criteria. We used design science with multiple iterations and let experts from different companies evaluate each cycle quantitatively and qualitatively. Finally, the prototype was implemented at a case company and the responsible team evaluated its efficiency.

Automated driving systems are safety-critical cyber-physical systems whose safety of the intended functionality (SOTIF) can not be assumed without proper argumentation based on appropriate evidences. Recent advances in standards and regulations on the safety of driving automation are therefore intensely concerned with demonstrating that the intended functionality of these systems does not introduce unreasonable risks to stakeholders. In this work, we critically analyze the ISO 21448 standard which contains requirements and guidance on how the SOTIF can be provably validated. Emphasis lies on developing a consistent terminology as a basis for the subsequent definition of a validation strategy when using quantitative acceptance criteria. In the broad picture, we aim to achieve a well-defined risk decomposition that enables rigorous, quantitative validation approaches for the SOTIF of automated driving systems.