What Auditors Really Look For: A Fireside Chat on ISO/PAS 8800 and Its Next Evolution
How do you audit the safety of artificial intelligence in road vehicles—and what comes after the first global automotive standard for AI safety?
In this Fireside Chat, SRES partners Jody Nelson, Gokul Krithivasan, and Bill Taylor are joined by Eduard Dojan, Product Manager for Software & Artificial Intelligence at SGS-TÜV Saar and a member of the ISO/PAS 8800 working committee, for an inside look at ISO/PAS 8800: Road Vehicles — Safety and Artificial Intelligence.
Together, they explore how AI safety is being audited in practice today, lessons emerging from early applications of the standard, and what development teams, safety engineers, and auditors should expect as ISO/PAS 8800 evolves toward its next edition. The discussion focuses on real-world challenges such as auditing non-deterministic behavior, dataset lifecycle, AI safety analyses, and integrating AI safety into existing functional safety and SOTIF frameworks.
Whether you’re developing ML-based ADAS or autonomous systems, supporting safety assurance, or preparing for future audits, this conversation will help you understand how AI safety is being operationalized—and where it’s headed next.
Watch the full discussion below:
Coming from LinkedIn mobile? Click here to watch on Youtube.
Selected Audience Q&A
The following questions were raised by attendees during the live session.
How is independence maintained between consulting and auditing when working with the same organization?
When organizations work together on projects, roles are separated so that consulting and auditing responsibilities remain truly independent. For instance, SRES supports consulting activities while SGS-TÜV Saar performs the audit independently.
When performing both activities within the same organization, the “I3” level of independence defined in ISO 26262-2:2018 must be implemented and assured within the organizational structures and policies if the vehicle-item is at ASIL D integrity; “I2” level for ASIL C.
Does ISO/PAS 8800 help AI usage in autonomous earth-moving vehicles?
The current scope for ISO/PAS 8800 is limited to on-road vehicles. ISO/IEC TR 5469 is a related industry-agnostic standard that can be a good reference. The ISO/IEC TR 5469 will soon be replaced by ISO/IEC TS 22440. When used in industries outside of automotive, it is possible to use ISO 8800 in combination with a different base functional safety standard like IEC 61508 or ISO 13849. In the fireside chat, Eduard confirms this approach as a viable route to getting certified in other industries using ISO/PAS 8800.
At SRES, we are already working with non-automotive clients by tailoring ISO/PAS 8800 concepts for their industry.
How does ISO/PAS 8800 expect organizations to make their safety criteria explicit? Does it require a defined harm boundary and acceptance logic up front, or does it assume those are inherited from other standards?
ISO/PAS 8800 assumes that vehicle- and system-level safety engineering work has already been performed using ISO 21448 (SOTIF) and ISO 26262. AI safety requirements are refined and allocated to the model from the “encompassing system” safety requirements.
As part of the SOTIF Hazard Identification and Risk Evaluation (ISO 21448, Clause 6), organizations define acceptance criteria using approaches such as GAMAB, ALARP, MEM or PRB. These acceptance criteria are expected to flow down to the AI system.
AI safety requirements are where model-level acceptance criteria or metrics—qualitative and quantitative—can be defined. ISO/PAS 8800 also introduces the concept of an “input space” derived from the Operational Design Domain (ODD) at the higher system level.
Meet the Speakers
Co-Founder & Managing Partner, SRES
Jody brings 20+ years of automotive safety experience in development, consulting, and audits. He approaches every project with the passion and care that his own family will ride in that vehicle. He began his career at Daimler AG in high voltage safety, EMC, software, and functional safety. In 2010 he co-founded kVA, and across kVA and now SRES, he has trained and supported hundreds of OEM, supplier, and semiconductor teams.
Co-Founder & Managing Partner, SRES
Gokul has 13+ years of experience in functional safety and AV safety, including leading development of robotaxis, commercial vehicles, and delivery robots. A certified expert in ISO 26262, ISO 21448, and ISO 8800, he has trained 1,000+ professionals and executives worldwide. At kVA, where he met Jody and Bill, he built global safety and cybersecurity teams across three continents; today at SRES he helps organizations deliver responsibly safe and secure products.
Partner, SRES
Bill brings decades of experience in functional safety, AV safety, and responsible AI. As co-founder of kVA, he helped set industry benchmarks for safety-critical development. At kVA and SRES, he has trained thousands of engineers, published over 20 technical papers, and earned awards from SAE and IEEE. He is dedicated to guiding organizations at the intersection of technology, safety, and responsibility.
Product Manager, Software & Artificial Intelligence — SGS-TÜV Saar
Eduard Dojan is Product Manager for Software & Artificial Intelligence at SGS-TÜV Saar and a member of the ISO/PAS 8800 working committee. He has over two decades of experience in automotive engineering and safety, including roles in functional safety and product development. In his current role, he supports AI and software assessment, auditing, and certification activities related to safety-critical systems.
About the Series
The idea for a series of Fireside Chats traces back to the very early days of Jody and Bill’s work together. In one of their first training sessions, a tricky FMEDA discussion left participants frustrated. That evening, the two of them stayed up late sketching a new way to explain it — an approach that ultimately found its way into the second edition of ISO 26262.
That moment sparked a tradition of technical discussions that has continued for years, now with the SRES team. These sessions have become a cornerstone of how we work — exploring difficult questions, challenging assumptions, and pushing toward better answers.
With these Fireside Chats, we’re opening that tradition to the broader safety community. Each session will bring the same candid, technical conversations that shape our work to a public forum.
Related Services
If you’re actively working on the topics discussed in this Fireside Chat, the following SRES services may be helpful.
- ISO 8800, AI-Safety Professional (AISP) Training – SGS-TÜV Saar
This certification training introduces the fundamentals of artificial intelligence and shows how established automotive safety standards can be supplemented with frameworks for responsible and safe AI, including ISO/PAS 8800. An optional exam is available to obtain the Artificial Intelligence Safety Practitioner (AISP) certificate.
- SRES Safety & Responsible AI Consulting
SRES is a consulting partner to OEMs, suppliers, AV developers, and EV manufacturers worldwide. Our experts support organizations in developing and validating the systems, processes, and safety cases needed to meet key automotive standards such as ISO 26262, ISO 21448 (SOTIF), ISO/SAE 21434, and ISO/PAS 8800.
