Definition
Systematically testing AI systems for vulnerabilities by exposing them to deliberately misleading or malicious input. Required for GPAI models with systemic risk. Includes red teaming, jailbreak attempts, and robustness testing.
Adversarial Testing
Definition & Explanation
Related Terms
Red Teaming
A testing method where experts try to "break" an AI system by finding vulnerabilities, bias, or unwa...
Systemic Risk
Risk that a GPAI model may have negative effects on public health, safety, public security, fundamen...
Robustness
The degree to which an AI system continues to function correctly under varying conditions, including...