Target discovery，a core step in drug development，is currently facing a dual bottleneck. Current research efforts remain highly concentrated on a small number of validated targets，resulting in serious homogeneity across drug pipelines. At the same time，traditional experimental and computational approaches are limited in throughput，precision，and their capacity to identify multiple targets in complex and refractory diseases. To overcome these challenges，artificial intelligence （AI） is now transforming this landscape. Natural language processing enables large-scale knowledge mining from biomedical literature，while the deep integration of multi-omics and multimodal datasets broadens biological insights. Convolutional neural networks extract features from bioimaging，and graph neural networks capture the complexity of molecular interaction networks. Together，these advances greatly accelerate the identification and validation of novel targets. Looking forward，the field is evolving toward AI-driven，multi-agent collaborative-optimization systems，offering a new paradigm for discovering “multi-target—multi-pathway” precision therapies. This paper systematically examines the innovative applications，challenges，and future directions of AI in target discovery，providing strategic guidance for the next generation of AI-enabled drug development.
Li et al. (Wed,) studied this question.