Adversarial Testing
Adversarial testing evaluates how an AI system behaves when exposed to intentionally malicious or misleading inputs.
“A test technique based on the attempted creation and execution of adversarial examples to identify defects in an ML model.”
(Definition reproduced from the ISTQB Glossary. Copyright belongs to ISTQB.)
View the complete ISTQB GlossaryWhat is Adversarial Testing?
Adversarial testing identifies weaknesses in artificial intelligence systems by deliberately attempting to influence or bypass expected model behavior using carefully designed inputs.
Security focus: Simulates intentional attacks against AI models.
Robustness evaluation: Determines how resistant a model is to manipulation.
Continuous improvement: Findings help strengthen AI reliability and security.
Real World Example
An online banking login flow is being reviewed after fraud analysts report repeated attempts to discover valid customer accounts.
The concern is not whether ordinary login works; it is how the system behaves when someone intentionally misuses it. Adversarial testing points testing toward that hostile or risky behavior.
The tester probes rate limits, error messages, audit events, lockout behavior, and monitoring signals to see whether the system gives attackers useful feedback.
The team strengthens the controls because Adversarial testing helps them test the product from the perspective of misuse, not only normal customer journeys.
Practice Questions
Question 1
What is the PRIMARY purpose of adversarial testing for ML models?
Question 2
Which of the following BEST distinguishes adversarial testing from fuzz testing?
Test your knowledge with real ISTQB-style questions
You’ve learned Adversarial Testing. Test your understanding with topic-specific questions in our Mock Exams.