MITRE ATLAS 2026 Top AI Detection: Adversarial Perturbation Crafting Against Inference Endpoint (AML.T0043)

Ibrahim Saud@tektrix

•updated Sep 25, 2026•0•0•0

Detects attempts to evade machine learning model classifications by crafting adversarial perturbations against an inference endpoint. The rule identifies suspicious patterns, including a high volume of near-identical inputs from a single client resulting in confidence scores near the model's decision boundary, which is indicative of an automated optimization loop searching for an adversarial evasion attack (MITRE ATLAS T0043).