Artificial Intelligence-Assisted Drug Delivery Systems: From Formulation Design to Personalized Therapeutics

Abstract

Drug delivery science has traditionally relied on iterative, trial-and-error experimentation to translate a promising molecule into a safe and effective dosage form. This approach, while proven, is slow, resource-intensive, and often struggles to account for the enormous variability between individual patients. Over the past decade, artificial intelligence (AI) and its subfields, machine learning (ML) and deep learning (DL), have begun to reshape this landscape by allowing formulators to predict, rather than merely observe, how a drug-excipient system, a nanocarrier, or an implantable device will behave in vitro and in vivo [1,2]. This review brings together current evidence on how AI is being applied across the full lifecycle of drug delivery systems (DDS), from early formulation screening and quality-by-design optimization to the engineering of nanocarriers, three-dimensional (3D)-printed personalized dosage forms, and closed-loop smart delivery devices. Particular attention is given to how these tools are converging with pharmacogenomics and real-time physiological monitoring to move the field toward genuinely individualized therapeutics. We also examine the practical obstacles that continue to limit widespread adoption, including fragmented and proprietary datasets, the limited interpretability of complex models, and an evolving but still uncertain regulatory environment. The review concludes that although AI is unlikely to replace experimental pharmaceutics in the near term, it is already functioning as a powerful hypothesis-generating and optimization partner, and its integration with automated experimentation and patient-derived data is likely to define the next generation of precision drug delivery.