Anfonwch hwn fel neges destun: Adversarial attacks can deceive AI systems, leading to misclassification or incorrect decisions