Artificial Intelligence in Drug Delivery for Infectious Diseases: A Critical Synthesis of Machine Learning Architectures, PBPK Modeling, and Antimicrobial Resistance

Objective: This paper provides a critical synthesis of the state-of-the-art AI methodologies applied to drug delivery systems for infectious diseases. This review aims to evaluate the integration of artificial intelligence (AI) and machine learning applications across four convergent pillars of drug delivery for infectious diseases: ML-guided drug formulation optimization; AI-driven targeted drug delivery system design; AI-augmented physiologically based PBPK modeling; ML and deep learning-based antimicrobial resistance profiling and management. Data Sources and Study Selection: A comprehensive systematic search of peer-reviewed literature was conducted using databases including PubMed, Scopus, IEEE and Google Scholar. Search strategies involve combination of keywords such as “machine learning”, “drug delivery”, “antimicrobial resistance”, “deep learning”, “infectious diseases”. Summary of the content of the article: The review aims to systematically map and categorize the principal AI and ML architectures according to their application and performance in drug delivery research. To critically assess the role of ML and DL based computational tools in drug formulation, AMR detection, susceptibility prediction, and resistance, their translational readiness, and contribution to broader AMR management strategies. To synthesize the current evidence on AI-integrated PBPK (Physiologically based Pharmacokinetic) modelling as a mechanism for improving the prediction of pharmacokinetic parameters in drug delivery systems. To identify the principal translational, regulatory, and methodological barriers to clinical deployment of AI-driven drug delivery tools for infectious diseases. Conclusion: Infectious diseases frequently induce rapid, dynamic pathophysiological shifts in the host, complicating precise drug dosing and increasing the risk of therapeutic failure due to antimicrobial resistance. Artificial intelligence represents the most cost-effective methodological advance available to infectious disease drug-delivery research. Deep neural networks and ensemble architectures consistently outperform classical computational models across drug optimization, nanoparticle delivery prediction, and AMR phenotype classification. Hybrid ML-PBPK frameworks achieves 65% accuracy as compared to 47.5% for conventional PBPK models. The transition from static to dynamic AI-enhanced PBPK modeling represents a vital advancement towards precision dosing, offering a robust paradigm to optimize therapeutic indices and improve patient survival rates in severe infections.