<![CDATA[The management of electric power grids is undergoing a significant transformation from conventional reactive approaches toward predictive and intelligent systems through the integration of Artificial Intelligence (AI). This transformation enables real-time data analysis, anomaly detection, automated fault diagnosis, and more accurate decision-making. This article aims to analyze the development, effectiveness, and implementation trends of AI-based electric power grid fault detection systems based on findings from previously published studies. A descriptive-comparative qualitative approach was employed through a systematic literature review of reputable scientific articles addressing the application of AI in fault detection, grid condition monitoring, equipment failure prediction, and decision-making in electric power systems. The analysis involved identifying major research themes, comparing AI methods, and synthesizing the advantages and limitations of each approach. The findings indicate that machine learning and deep learning techniques, particularly when integrated with Supervisory Control and Data Acquisition (SCADA) systems and the Internet of Things (IoT), can significantly improve fault detection accuracy, accelerate fault localization, reduce outage duration, and enhance the reliability and resilience of electric power systems. AI integration also facilitates predictive maintenance, real-time operational decision-making, and the development of self-healing smart grids. However, challenges remain regarding data quality, cybersecurity, system interoperability, computational requirements, and model interpretability. Future research should therefore focus on developing adaptive, transparent, and robust AI models capable of operating effectively in complex and dynamic power grid environments.]]>
