From Data‐Driven to Intelligence‐Driven: Innovating Research Paradigms of Perovskite Solar Cells

Perovskite solar cells (PSCs) are regarded as the next‐generation photovoltaic technology due to exceptional power conversion efficiency (PCE) and low fabrication costs. However, their multi‐scale physical and chemical structures render empirically driven research paradigms inefficient, especially like trial‐and‐error approach. In recent years, classical machine learning methods based on big data have been widely applied in PSCs research, yet their effectiveness remains constrained by issues such as limited data quality, single modality of data, and insufficient generative capabilities. With the rapid advancement of artificial intelligence, particularly the rise of deep learning, reinforcement learning, and large language models, PSCs' research progressively transforms from a local data‐driven to a global intelligence‐driven paradigm. This review systematically summarizes recent progress in intelligence‐driven methods for PSCs research, encompassing material design, process optimization, exploration of stability mechanisms, knowledge discovery, and self‐driving experimentation. We further outline key points and current challenges of AI technologies in enhancing PCE, stability, and scalability of PSCs, aiming to accelerate the development from laboratory innovation to industrial application.