Adversarial Machine Learning on AI Model Attacks