ISTQB® CT-AI certifies testing of AI-based systems (ML + generative AI/LLMs). Syllabus v2.0 is current; v1.0 English exams run until 21 Apr 2027 (non-English until 21 Oct 2027).
Exam: 40 MCQs, 60 min, 44 points (pass 29 ≈ 66%). Higher K-level questions score >1 point; non-native speakers get +25% time. Prerequisite: CTFL.
It covers the real challenges of AI testing—probabilistic behaviour, non-determinism, data dependence—across seven areas including quality characteristics (ISO/IEC 25059), input data testing, model testing and ML development testing.
Aimed at testers, test managers, data scientists and developers. Delivered year-round by accredited providers (fee set locally, e.g. Indian Testing Board in India). Do not confuse with CT-GenAI (which is about *using* GenAI for testing).
Format
Computer-based (test centre or remotely proctored), delivered on demand year-round by accredited exam providers. No fixed national exam date.
Questions
40
Marking
40 questions worth 44 points in total — higher K-level questions carry more than one point. Pass mark 29 of 44 (approximately 66%). No negative marking.