Applying software automotive engineering practices beyond the automotive industry

Introduction

The automotive industry has developed rigorous approaches for dealing with safety in increasingly complex software-driven systems. Standards such as ISO 26262 and ISO 21448 (SOTIF) provide structured methods for identifying risks, analyzing system limitations, and systematically validating intended behavior.

In this article, I explore how these practices can be applied beyond the automotive domain. In particular, I discuss Hazard Identification and Risk Evaluation (HIRE), a technique used within SOTIF to identify potentially hazardous scenarios, evaluate their consequences, and define appropriate mitigation and verification measures.

The same way of thinking can be valuable for other domains where software and AI systems operate in complex or high-stakes environments. The article illustrates this idea through the example of AI-based tools used in legal workflows, where functional limitations and insufficient verification can have significant consequences.

The goal is not to directly transfer automotive standards to other industries, but to adapt the engineering principles behind them: systematic risk identification, scenario analysis, verification, and continuous reduction of unknown and potentially hazardous situations.

The article was published in NoticIEEEro #132: Read the full article in NoticIEEEro #132