Bridging artificial intelligence with photovoltaics

Turning the significant possibilities presented by artificial intelligence (AI) into practical outcomes for sustainable-energy technology advancement remains challenging, as it requires bridging the independently developed energy and AI domains. This review leverages insights from the photovoltaics (PV) field to highlight the opportunities and challenges presented by AI while simultaneously providing an organized overview to AI researchers on the status of AI application in the PV domain. The current AI application landscape is classified according to the tasks and features required at each stage of PV development. The five major domains identified are material development, process development, fault diagnosis, PV power forecasting and planning, and operation and management. The status and challenges of AI applications are provided, highlighting imbalance in research fields, data dependency, the relationships between components of PV devices, and energy consumption of AI. Further advancements will hinge on our ability to enable AI to operate safely beyond existing datasets so that we can unlock its potential to explore uncharted territories in sustainable technologies and achieve a transformative impact. Lee et al. review how artificial intelligence (AI) connects with photovoltaics (PV). They identify five key domains—materials discovery, process optimization, fault diagnosis, power forecasting, and system management—and highlight challenges such as dependence on human-made data while outlining how AI can move beyond existing datasets to unlock future opportunities.