AI-driven drug design: a comprehensive review

The integration of AI into drug design has undergone a transformative evolution, reshaping the landscape of medicinal chemistry. AI's exceptional capabilities in data processing, pattern recognition, and predictive modelling have permeated every stage of the drug development pipeline. To provide a comprehensive overview of the current state, key methodologies, and emerging trends in this rapidly evolving field, this study systematically examines global research achievements in AI-driven drug design and discovery over the past 22 years. Drawing upon the Science Citation Index-Expanded and Social Sciences Citation Index databases, a multi-dimension analysis was conducted on AI-driven drug design spanning from 2004 to 2025. The dataset underwent rigorous cleaning, knowledge discovery, and visualization using the Derwent Data Analyzer. 16,190 publications in total were systematically reviewed. The findings reveal that China, USA, and India are the most prolific contributors to AI-driven drug design research, with USA demonstrating the highest citation volume. The Chinese Academy of Sciences ranked first in both publication output and H-index, while the Harvard University exhibited the highest average citation per publication. Journal of Chemical Information and Modeling emerged as the most productive journal, and "CHEMISTRY, MULTIDISCIPLINARY" was the predominant disciplinary category. Current research focus includes protein structure prediction, deep learning, drug repurposing (or drug repositioning), and artificial neural networks. Additionally, emerging research frontiers such as chemical language models for network-based target screening, AI in clinical trials, predictive toxicology, and adverse drug reaction analysis are anticipated to drive innovation in the coming years.