The International Software Testing Qualifications Board (ISTQB®) has released the Certified Tester AI Testing (CT-AI) Syllabus Version 2.0, marking a significant update to its specialist certification in AI testing.
As AI systems move into production across a wide range of industries, the expectations around quality, reliability, and trust continue to rise. Version 2.0 reflects this shift, with a stronger focus on how AI-based systems are tested in practice, particularly those built on machine learning and generative AI.
A more focused and streamlined syllabus
The new syllabus has been restructured into a clearer and more cohesive learning path. The content is now organized around the key elements of AI systems, including data, machine learning models, and system-level testing. This structure better reflects how AI systems are developed and tested in real-world environments.
The recommended training duration has been reduced from four days to three, while maintaining the depth of coverage and improving practical relevance.
A clear focus on testing AI-based systems
Version 2.0 concentrates entirely on testing AI-based systems. Content related to using AI for testing has been removed, allowing the syllabus to go deeper into the specific challenges that arise when testing systems based on machine learning.
These challenges include probabilistic behavior, the difficulty of defining test oracles, and the need for statistical approaches to testing.
Expanded coverage of generative AI and large language models
The syllabus now includes dedicated coverage of testing generative AI and large language models. This includes practical techniques such as exploratory testing and red teaming, reflecting how these systems are currently assessed for quality, robustness, and potential misuse.
Comprehensive coverage of the AI testing lifecycle
Version 2.0 introduces a structured approach to testing across the machine learning lifecycle, including:
- Input data testing, covering bias, data representativeness, label correctness, and data pipeline validation
- Model testing, including adversarial testing, metamorphic testing, drift testing, A/B testing, and back-to-back testing
- AI-specific quality characteristics, aligned with standards such as ISO/IEC 25059
- Risk-based and statistical testing approaches suited to machine learning systems
These areas are supported by hands-on exercises that help candidates apply the concepts in practical scenarios.
Stronger emphasis on practical skills
The updated syllabus includes hands-on exercises across key topics, including machine learning workflows, data preparation, performance evaluation, and testing of generative AI systems. This ensures that candidates gain both practical experience and theoretical understanding.
Responding to the needs of modern AI systems
“AI-based systems introduce new and complex testing challenges,” said Klaudia Dussa-Zieger, President of ISTQB® and Chair of the CT-AI Taskforce. “Version 2.0 provides a clearer and more practical foundation for testers working with machine learning and generative AI technologies. It reflects the way AI systems are evolving and the growing need for structured testing approaches”.
Who should consider ISTQB® CT-AI certification?
The CT-AI certification is designed for professionals involved in testing AI-based systems, including testers, test analysts, test engineers, developers, data professionals, and test managers. It is also suitable for anyone seeking a structured understanding of AI testing challenges.
Further information
More information about CT-AI Version 2.0 including the syllabus, sample exam, and accreditation guidelines, is available here, and frequently asked questions (FAQs) can be found here.