AITG-MOD-02 control for runtime model poisoning testing, verifying the resilience of the AI model against attacks that attempt to corrupt or manipulate the model during execution, through malicious inputs or adversarial machine learning techniques.