Abstract: Space weather prediction remains a grand challenge due to the nonlinear, multiscale coupling between solar activity, the heliosphere, and Earth’s magnetosphere-ionosphere-thermosphere system. Over the past decade, artificial intelligence (AI) and machine learning (ML) have emerged as powerful complements to physics-based models, offering improved forecast skill, reduced latency, and probabilistic decision support. This review synthesizes recent advances in AI-driven space weather prediction, critically evaluates benchmark performance across forecasting tasks (solar flares, coronal mass ejections, geomagnetic storms, and ionospheric disturbances), and discusses hybrid AI–physics frameworks, explainability, operational readiness, and governance challenges. We argue that AI is reshaping space weather forecasting from empirical pattern recognition toward autonomous, uncertainty-aware prediction systems.

Keywords: Space weather, artificial intelligence, machine learning, geomagnetic storms, solar flares, digital twins

Pages: 149-166

WSEAS Transactions on