Basic safety and Robustness: AI systems may be at risk of adversarial attacks, the place modest, imperceptible variations to the input can cause the process to make faulty or even hazardous decisions. Guaranteeing the protection and robustness of AI techniques is actually a crucial obstacle.