What is an Adversarial Distillation Attack?

An Adversarial Distillation Attack is when an attacker deliberately interferes with the process of training a smaller AI using a larger, more powerful one. By manipulating what the smaller model learns, the attacker can cause it to make specific mistakes, develop security weaknesses, or behave in misleading or harmful ways. The resulting model may appear to work normally, making the damage difficult to detect.