Artificial intelligence has been increasingly adopted across power and energy systems to address challenges related to variability, uncertainty, and operational complexity. In recent years, machine learning and deep learning techniques have been applied to a wide range of tasks, including renewable energy forecasting, energy management, grid monitoring, and electricity market analysis. Rather than treating these applications in isolation, this review examines how different AI method families are utilized across distinct energy system layers and decision-making contexts. Emphasis is placed on the practical implications of AI-based methods for system operation, reliability, and planning, as reported in the existing literature. By synthesizing findings across forecasting, control, cyber–physical monitoring, and market applications, this study aims to clarify both the capabilities and current limitations of AI-driven approaches in modern power systems. Across 2015–2025 publications, short-term PV power forecasting models achieve typical median MAPE in the range of 2%–5% over dozens of benchmark datasets, while deep reinforcement learning-based microgrid energy management systems report operational expenditure reductions on the order of 8%–15% compared to rule-based baselines. The survey covers advances in signal decomposition (DWT, EWT, VMD), hybrid and probabilistic models, cyber–physical security, and emerging edge-AI deployments. • Comprehensive review of AI applications in modern energy systems. • AI-based optimization for efficiency and renewable integration. • Key challenges in intelligent energy management and control. • Future research directions for sustainable, resilient power grids. • Interdisciplinary roadmap connecting AI with sustainable power systems.
Dzafic et al. (Wed,) studied this question.