The intersection of artificial intelligence and cybersecurity represents one of the most prominent and sensitive fields in modern technology. Artificial intelligence has become a dual-purpose tool that can be employed to strengthen defensive capabilities or exploited to carry out cyberattacks. On the defensive side, machine learning and deep learning algorithms contribute to analyzing vast amounts of system logs and network data, helping anticipate threats, accurately identify abnormal behavioral patterns, automate security incident response procedures, and reduce Mean Time to Detect and Mean Time to Respond (MTTD/MTTR). Conversely, malicious actors exploit artificial intelligence technologies to launch more sophisticated and targeted cyberattacks. These include advanced phishing campaigns generated using generative AI, the development of malware capable of modifying its code to evade detection, and Adversarial Attacks designed to mislead machine learning models themselves. This ongoing technological arms race highlights the importance of developing defensive systems based on Explainable Artificial Intelligence (XAI), thereby enhancing the reliability of security decisions and supporting the development of a more efficient cybersecurity framework capable of confronting evolving threats.