Artificial Intelligence in New Drug Development: Current Applications and Future Prospects

Abstract

To systematically evaluate the current applications, key technologies, and implementation outcomes of artificial intelligence (AI) across the entire drug development process, analyze existing challenges, and explore future directions. Methods: Literature published between 2020 and 2025 was retrieved from core databases including PubMed, MEDLINE, Embase, and the Cochrane Library, focusing on AI applications in target identification, compound design, preclinical research, and clinical trials. A systematic review approach was adopted to summarize implementation pathways, performance metrics, and comparative results versus traditional methods. Results: Over 80 eligible studies were included. Findings indicate that AI demonstrates significant advantages at all stages of drug development: in target identification, AI improved efficiency by approximately 40%–60%; in compound design, generative models reduced lead compound discovery time to one-third to one-half of the original duration; in clinical trials, AI-assisted patient stratification increased recruitment efficiency by about 30%–50%. Additionally, AI models achieved 85%–92% accuracy in predicting drug toxicity and pharmacokinetic parameters, markedly outperforming conventional computational methods. Conclusion: AI is profoundly transforming the paradigm of drug development, leveraging data-driven approaches to substantially enhance efficiency, reduce costs, and improve success rates. Despite challenges in data quality, algorithm interpretability, and regulatory compliance, ongoing technological innovation and multidisciplinary integration are expected to drive drug development into a more precise and efficient era.