Advanced Level – Cross-Cutting
Machine Learning Testing – ISTQB Definition
Last Updated: July 12, 2026 · SoftwareTestPilot QA Team
Official ISTQB Definition
Testing of systems that use machine learning models, covering data quality, model accuracy, fairness, robustness and drift over time.
In simple words
Test ML systems on data, model behavior, fairness and drift — not just standard functional tests.
Exam tip
ML testing needs metamorphic and property-based tests because exact expected outputs are usually unknown.
Related terms
- Metamorphic TestingInstead of asserting an exact answer, assert that changing the input in a known way changes the output in a known way.
- Property-Based TestingInstead of writing example tests, describe rules that should always hold and let the tool try 1000s of inputs.
- AI-Based TestingUsing AI/ML in the test process itself — auto-generating tests, self-healing selectors, predicting risky areas.
Practice this term in the ISTQB Mock Test
40 CTFL v4.0 questions, 60-minute timer, instant scoring and chapter-wise breakdown.
Practice in Mock Test