Announcing the ISO/IEC 42119 Series: A New Era for AI Testing & Assurance

We are excited to announce the arrival of the ISO/IEC 42119 series – a groundbreaking set of international standards dedicated to the testing of artificial intelligence (AI) systems. This series is set to become a cornerstone for organizations seeking robust, transparent and trustworthy AI, complementing the foundational ISO/IEC 42001 standard for AI management systems.

Celebrating the first milestone: ISO/IEC TS 42119-2:2025

The journey begins with the publication of ISO/IEC TS 42119-2:2025 – Overview of Testing AI Systems. This provides a comprehensive introduction to the concepts, processes and test approaches for AI, showing how the established ISO/IEC/IEEE 29119 software testing standards can be applied in the AI context. It covers:

• The AI system life cycle and risk-based testing
• Test processes and documentation tailored for AI
• Test levels and types, including data quality and model testing
• Guidance on identifying and treating risks unique to AI systems
• How to leverage established test design techniques and coverage measures for AI

While this first part is an overview, it marks a significant step forward – providing organizations with a structured approach to AI testing and laying the groundwork for more specialized guidance to come.

What’s next?

The ISO/IEC 42119 series is just beginning. Here is a summary of what is on the horizon:

• ISO/IEC DTS 42119-3 – Verification and Validation Analysis of AI Systems (Draft):
• Will describe approaches for verification and validation (V&V) analysis of AI systems, including formal methods, simulation and evaluation across the AI life cycle. The focus is on providing strong evidence that AI systems behave as intended
• ISO/IEC AWI TS 42119-7 – Red Teaming (Under Development):
• Will codify structured adversarial testing (red teaming) to probe the robustness, security and misuse risks of AI systems – especially those that are safety-critical or high-impact
• ISO/IEC AWI TS 42119-8 – Quality Assessment of Prompt-Based Text-to-Text GenAI Systems (Under Development):
• Will focus on large language model (LLM)-based, prompt-driven systems, combining quality assessment (including safety) with methods such as red teaming and offering guidance on analyzing and documenting results

Each new part will address specific challenges and best practices for testing and assuring different types of AI systems, from traditional expert systems to cutting-edge generative AI.

How does ISO/IEC 42119 support AI assurance?

The ISO/IEC 42119 series is designed to work hand-in-hand with ISO/IEC 42001, the world’s first AI management system standard. While ISO/IEC 42001 sets out requirements for responsible AI governance, ISO/IEC 42119 provides the technical foundation for testing and validating AI systems – ensuring they are reliable, safe and trustworthy.

Key benefits include:

• Risk-based testing: Aligns test strategies with the unique risks of AI, supporting compliance and stakeholder trust
• Life cycle coverage: Addresses testing needs across the entire AI system life cycle, from data and model development to deployment and monitoring
• Integration with existing standards: Builds on the ISO/IEC/IEEE 29119 series, enabling organizations to leverage familiar processes and documentation in the AI context
• Support for emerging AI technologies: Upcoming parts will address the latest developments, including generative AI and adversarial testing

AI assurance service offerings

We are a global leader in AI assurance, digital trust, testing and certification services. We provide a comprehensive portfolio to help organizations navigate the evolving landscape of AI standards, regulations, and best practices:

• Certification: Accredited ISO/IEC 42001 AI Management System certification to ensure governance and compliance, along with other recognized certification schemes
• Advisory and training: Expert guidance on AI risk management, lifecycle assurance, and readiness for global standards and regulations. SGS Academy offers courses on AI standards, the EU AI Act, risk management, and technical topics
• Testing: AI system testing aligned with ISO/IEC 42119 and ISO/IEC/IEEE 29119, including:
  • Data quality, robustness, fairness, explainability, and adversarial testing
  • Technical product testing for LLMs, chatbots, and other GenAI systems
  • Red teaming and security assessments
• Guidepost for Trustworthy AI: Start with the SGS AI Trust Check – a rapid self-assessment tool that helps identify gaps and prioritize actions for compliance and trustworthiness

Why choose SGS?

• Global reach: 500+ digital trust experts, 10 accredited cyberlabs and a presence in over 140 countries
• Pioneering leadership: First to deliver accredited AI management system certification and end-to-end EU AI Act compliance services
• Comprehensive portfolio: Covering certification, testing, training, advisory and digital trust solutions for every stage of your AI journey
• Industry expertise: Serving clients across finance, healthcare, mobility, manufacturing and more

Ready to start your AI assurance journey?

We are committed to supporting clients on their AI assurance journey – helping you navigate new standards, implement best practices, and achieve compliance with confidence.

Contact us today to learn how we can help you leverage the ISO/IEC 42119 series and build trustworthy, responsible AI.

For further information, please contact:

Michal Cichocki
Global Product Manager – AI Assurance Services
Business Assurance
t: +48 22 329 22 22