integrity violation

In the AML context, an AI system being forced to misperform against its intended objectives, producing outputs or predictions that align with the at­tacker’s objective.