Understanding ISO/IEC TS 22440
Artificial Intelligence – Functional Safety and AI Systems
Overview
Artificial intelligence (AI) and machine learning (ML) have the potential of becoming core components in safety-critical systems. However, traditional functional safety standards, like IEC 61508, are not currently designed to address the non-deterministic, adaptive, and data-driven nature of AI.
Reynolds & Moore is hosting a comprehensive 2-day training, instructed by RMAI, designed to provide a deep understanding of designing safety systems using AI/ML in accordance with ISO/IEC TS 22440: Artificial Intelligence – Functional Safety and AI Systems. The training will reflect the latest updates to the technical specification supporting the safe integration of AI/ML into systems where human safety, critical operations, or compliance are essential.
This training is ideal for engineers, project managers, and safety professionals intending to leverage AI/ML in their safety systems.
Why It Matters
ISO/IEC TS 22440 is the first international standard developed specifically to guide development, verification, and validation of AI-enabled functional safety for industrial applications. Without a defined safety framework, AI systems, despite their potential, face serious hurdles in acceptance, certification, and deployment in regulated industries. ISO/IEC TS 22440 provides the methodology to analyze, assess, and mitigate AI-related risks, enabling the development of:
- AI-based behavioral decision-making in safety critical systems
- AI-enabled identification and classification of objects
- Predictive diagnostics in industrial maintenance
- Machine learning models used in safety monitoring or anomaly detection
This standard equips organizations with the tools to confidently deploy AI in safety-critical decision-making, enabling responsible innovation while ensuring compliance and trust.
Course Structure
Day 1: In-depth Review of ISO/IEC TS 22440-1 & ISO/IEC TS 22440-2
- Introduction to ISO/IEC TS 22440 and its scope
- Three-stage analysis for the realization of AI systems:
- Data acquisition
- Knowledge induction
- Processing and output generation
- Classification of AI technology to support identification of risk reduction and mitigation methods
- AI safety lifecycle within the traditional functional safety lifecycle
- Unique hazard and risk assessment considerations for AI systems
- Failure modes and affects specific to AI systems
- Risk reduction methods and mitigations to be considered for AI systems
Day 2: Examples of ISO/IEC TS 22440 Applications & Exam
- Wrapping up ISO/IEC TS 22440-1: Verification and validation requirements
- Review of application examples from ISO/IEC TS 22440-3
- Open discussion for remaining questions
- 2-hour open notes exam
This course is designed to be interactive and practical, with opportunities for discussion, Q&A, and application of concepts to real-world scenarios. By the end of the training, participants have a solid foundation in both theories and applications of AI functional safety principles. The exam requires a passing score of a minimum of 70%. Upon successful completion of the exam, participants will receive a Confirmation of Participation of the training.
Enrollment
The participation cost is $3,000 and enrollment is limited to 20 participants to ensure a high-quality learning experience. Seats are allocated on a first-come, first-served basis.
Disclaimer: This training is based on a draft of ISO/IEC TS 22440 and reflects the latest available updates to the technical specification. As the standard continues to evolve, content may be updated accordingly. The aim is to provide a timely and practical understanding to support the safe and effective integration of AI/ML into safety-critical systems.
