Purpose of Review <p>This review aims to systematically synthesize recent developments in artificial intelligence (AI)-driven power system planning under the low-carbon transition. It focuses on how AI methodologies address increasing system complexity and uncertainty by improving scenario generation, model optimization, and decision support transparency.</p> Recent Findings <p>Recent advances demonstrate that deep learning, reinforcement learning, and generative models significantly enhance input accuracy and optimization efficiency for large-scale, non-convex planning problems. Explainable AI techniques, such as SHAP and LIME, have been increasingly integrated to improve the interpretability and credibility of planning outcomes. However, challenges persist in ensuring physical consistency in scenario generation, modelling realistic multi-agent interactions, and developing trustworthy AI frameworks.</p> Summary <p>AI technologies are reshaping power system planning by offering intelligent, robust, and sustainable solutions aligned with high renewable penetration and evolving market dynamics. Future research must address key gaps in physical feasibility, behavioural realism, and explainability to fully leverage AI’s potential for supporting low-carbon power system transitions.</p>