Definition
A systematic, adversarial testing process where a dedicated team attempts to find flaws, vulnerabilities, and potential harmful biases in an AI system before its deployment.
In short: A systematic, adversarial testing process where a dedicated team attempts to find flaws, vulnerabilities, and potential harmful biases in an AI system before its deployment.