The presented work presents the main security risks in artificial intelligence and the defense mechanisms used to prevent them. Data protection is necessary for the establishment of safe and reliable artificial intelligence and for artificial intelligence systems to serve the development of society in a reliable, transparent and responsible manner. Cybersecurity methods and risks in artificial intelligence are reflected here. This work proposes a cybersecurity-aware artificial intelligence (AI) algorithm that integrates data security, threat detection, secure model training, and continuous monitoring.