Large Models as a New Quality Medical Technology Trigger the Methodological Evolution of Drug Research and Development

By analyzing cutting-edge cases such as the AlphaFold series and Insilico Medicine's Pharma.AI platform, and by integrating actor-network theory from the philosophy of scientific practice, this study examines the logical reconstruction across target discovery, molecular generation, and clinical prediction. While traditional drug research and development relies heavily on the linear logic of "hypothesis-experiment-verification" and the physical laboratory setting, the intervention of artificial intelligence is shifting scientific knowledge toward a process of prediction, generation, and reorganization within algorithmic networks. Moving beyond mere "instrumental assistance", artificial intelligence—through pattern recognition, hypothesis generation, and model-based deduction—demonstrates an agency distinct from typical non-human actors, effectively empowering research and development productivity. The integration of artificial intelligence is not only reshaping the technical pathways of drug discovery but also restructuring the actor-network of scientific practice itself.