AI-driven power system flexibility: a review of demand response, renewable integration, and grid reliability

The global energy sector is undergoing a significant transformation, driven by the need for deep decarbonization and the rapid advancement of artificial intelligence (AI). This review explores the integration of AI-driven flexibility, demand response (DR), and renewable energy, aiming to achieve a secure and sustainable energy transition. We present a systematic analysis of high-quality research publications, emphasizing recent progress in AI applications for power systems. A structured taxonomy of AI techniques tailored for power system flexibility is established, addressing critical challenges like renewable intermittency and grid stability. Influential AI methodologies deployed across power systems are discussed, alongside a critical analysis evaluating their techno-economic trade-offs and implementation readiness. Our review offers an integrated analysis of AI-driven grid applications that addresses both technical performance and operational challenges, culminating in a proposed roadmap for promising interdisciplinary approaches. The evidence reviewed indicates that AI-driven flexibility reduces integration costs for variable renewables, significantly improves DR participation rates, and enhances grid reliability beyond conventional methods. This synthesis serves as a coherent roadmap for researchers and practitioners working on intelligent energy systems.