The use of AI/ML is rapidly expanding across safety-critical sectors, including automotive. However, its non-deterministic nature challenges traditional functional safety frameworks like ISO 26262 and IEC 61508, which rely on traceability and predictable behavior. Emerging standards such as ISO/IEC TR 29119-11, ISO/IEC DTS 25058, ISO PAS 8800, and ISO DTS 5083 aim to address these gaps. Until then, ensuring the reliability, safety and security of AI-driven systems remains a complex task. This presentation explores how the industry can balance the need for innovation with the demands of safety assurance in AI/ML-based automotive applications.

• The challenges AI/ML introduces to ISO 26262 and functional safety frameworks in automotive systems
• Emerging AI/ML safety standards like ISO/IEC TR 29119-11, ISO PAS 8800 and ISO DTS 5083
• The conflict between AI’s non-determinism and the need for traceability and predictability in safety-critical systems
• Insights into strategies for assuring safety, reliability and security in AI-driven automotive applications
• How AI/ML can enhance safety beyond traditional development methods while maintaining compliance expectations