Threat intelligence / Mitigation
Validate AI Model
Validate that AI models perform as intended by testing for backdoor triggers, potential for data leakage, or adversarial influence. Monitor AI model for concept drift and training data drift, which may indicate data tampering and poisoning.
Mitigation
TypeMitigation
External ID—
First seen2026-08-14
Last seen2026-08-14
Confidence
—
Validate that AI models perform as intended by testing for backdoor triggers, potential for data leakage, or adversarial influence. Monitor AI model for concept drift and training data drift, which may indicate data tampering and poisoning.
10 recorded
Relationships
1 external