Advancing perovskite photovoltaic technology through machine learning‐driven automation

Since its emergence in 2009, perovskite photovoltaic technology has achieved remarkable progress, with efficiencies soaring from 3.8% to over 26%. Despite these advancements, challenges such as long-term material and device stability remain. Addressing these challenges requires reproducible, user-independent laboratory processes and intelligent experimental preselection. Traditional trial-and-error methods and manual analysis are inefficient and urgently need advanced strategies. Automated acceleration platforms have transformed this field by improving efficiency, minimizing errors, and ensuring consistency. This review summarizes recent developments in machine learning-driven automation for perovskite photovoltaics, with a focus on its application in new transport material discovery, composition screening, and device preparation optimization. Furthermore, the review introduces the concept of the self-driven Autonomous Material and Device Acceleration Platforms (AMADAP) laboratory and discusses potential challenges it may face. This approach streamlines the entire process, from material discovery to device performance improvement, ultimately accelerating the development of emerging photovoltaic technologies.