AI-driven fungicide design: From target identification to field application

Plant pathogenic fungi severely threaten global agriculture, causing substantial yield losses in staple crops and endangering food safety through mycotoxin contamination. Conventional fungicide development suffers from high costs, lengthy timelines, and rapid evolution of fungal resistance that outpaces traditional discovery workflows. While artificial intelligence (AI) offers transformative potential to address these bottlenecks, its application in plant pathology remains fragmented, lacking integration of agricultural-specific constraints such as field stability, ecological safety, and resistance management. This review presents the AI-driven fungicide design (AIFD) platform, a comprehensive framework comprising four interdependent components: a plant pathogen-specific data ecosystem, a modular microservice technical architecture, a linear multi-phased development workflow, and a specialized resistance prediction workflow. We synthesize key technological progress across the fungicide development pipeline, from target identification and virtual screening to molecular optimization and field validation, emphasizing AI methodologies adapted to agrochemical requirements rather than pharmaceutical standards. Despite significant advances, critical challenges persist, including scarce high-quality training data for understudied pathogens, limited model adaptability across diverse agroecosystems, poor interpretability hindering stakeholder trust, and accessibility barriers for resource-constrained researchers. Future directions emphasize real-time field data integration, explainable AI for regulatory acceptance, and inclusive design strategies to bridge the lab-to-field gap. By aligning computational innovation with agricultural priorities, AIFD platforms accelerate the discovery of resistance-breaking, environmentally benign fungicides, offering a viable pathway toward sustainable crop protection and enhanced global food security.