Prof. Dr. Mirko Conrad
Managing Director | samoconsult GmbH
More Info
ISO/PAS 8800 is becoming a cornerstone for automotive organizations developing or integrating AI systems, because it defines how to manage AI‑related risks in a structured and industry‑aligned way. As AI moves into safety‑critical applications, development teams must understand how to identify AI‑specific hazards, establish trustworthy development practices, and demonstrate compliance to customers and regulators.
Gaining knowledge about ISO/PAS 8800 equips engineers, managers, and safety professionals with a common framework for responsible AI development and prepares organizations for upcoming regulatory expectations and future ISO standards built on this PAS.
This one-day AI safety training class introduces participants to the core concepts of AI safety, covers the scope, intent, and structure of ISO/PAS 8800, discusses AI‑specific risk sources and the lifecycle activities required to develop and operate trustworthy AI systems. Practical examples help participants understand how to ensure AI safety in real development environments and how to prepare their organization for compliance with ISO/PAS 8800.
Target Audience
This training is designed for automotive professionals involved in the development, verification, and validation of AI‑enabled systems, including engineers, data scientists, safety managers, product owners, and technical leads. It is equally relevant for safety auditors/assessors, quality managers, project managers, and decision‑makers who need to understand AI‑specific risks and the requirements of ISO/PAS 8800.
Highlights
- AI concepts and terminology
- ISO/PAS 8800 scope, intent, and structure
- Reference AI safety lifecycle
- Safety‑related properties of AI systems
- AI errors and root causes
- Dataset considerations
- Practical examples
Agenda
- Artificial Intelligence
- Machine Learning
- GenAI
- Scope and intent
- Structure
- Conceptual framework
- Activities
- Work products
AI errors and their root causes
- Observation, model, label, operation
Derivation of AI safety requirements
Safety-related properties of AI systems
- Dataset lifecycle
- Dataset-related safety properties
Architectural and development measures (incl. redundancy concepts for ML systems)
Verification, validation, and safety analysis of AI systems
Assurance argumentation for AI systems
Confidence in the use of AI tools/frameworks
AI safety management
Our Trainer
Available as in-house training class
- Exclusively for your team or organization.
- We adapt the content, starting times, and agenda to your specific needs and challenges.
- Delivered either online or in person at your organization.
