Integrating Artificial Intelligence in the Development Process of Functional Safety according to ISO 26262

Safety-critical system development remains one of the most demanding areas of engineering, especially in the automotive domain, where system failures can have severe consequences. The increasing complexity of electronic and software-based systems has intensified the need for structured, verifiable development processes to meet required safety levels. ISO 26262 provides such a framework for achieving functional safety throughout the entire automotive product lifecycle. In this context, the rapid progress of Artificial Intelligence (AI) introduces new opportunities but also challenges that existing safety standards have not yet thoroughly addressed. This contribution investigates the potential use of AI supporting tools within the development process of safety-relevant control functions. Based on a literature review and a comparative analysis of normative and regulatory frameworks, including ISO 26262, IEC 61508, ISO/PAS 8800, and the EU AI Act, potential AI tools are identified for supporting hazard and risk analysis (HARA), requirements engineering, definition of system and software architecture, and verification and validation activities. Based on practical examples, the paper explores how AI-based methods such as Natural Language Processing and Machine Learning could support information extraction, classification, and consistency checking while maintaining normative compliance and process transparency. The results indicate that AI can conceptually enhance efficiency and assist engineers in the safety development process, especially when handling large and complex data sets, such as in requirements engineering and documentation activities. Nevertheless, there still exist challenges regarding limited explainability, reproducibility, and the absence of clear qualification procedures. In conclusion, this paper contributes to a structured understanding of how AI can be aligned with existing safety frameworks, and it underlines that the integration of AI should follow a gradual and well-defined process, supported by clear verification methods, human supervision, and regular evaluation to maintain safety, normative compliance, and trust in safety-critical system development.

Keywords: Artificial Intelligence, Road Vehicles, ISO 26262, Functional Safety, V-Model Development Process.