Recent Advances in AI-Assisted Drug Design and Discovery: A Review
Keywords:
Artificial intelligence, Drug discovery, Generative AI, Molecular design, Virtual screening, Machine learningAbstract
Artificial intelligence (AI) has emerged as a transformative technology in pharmaceutical research, particularly in drug design and discovery. Conventional drug discovery is a complex, expensive and time-consuming process involving target identification, hit discovery, lead optimization, preclinical evaluation and clinical development. The increasing availability of chemical, biological and clinical datasets, together with advances in machine learning and deep learning, has created new opportunities for accelerating several stages of this process. AI-assisted approaches are currently being applied to target identification, virtual screening, molecular property prediction, quantitative structure–activity relationship analysis, de novo molecular design, drug repurposing, synthesis planning and protein-structure prediction. Generative artificial intelligence has further expanded these capabilities by enabling the design of novel molecular structures with desired physicochemical and biological characteristics. Transformers, variational autoencoders, generative adversarial networks, reinforcement learning and diffusion models have emerged as important computational approaches for generative molecular design. Recent developments such as AlphaFold 3 have expanded AI-based structural prediction toward complexes involving proteins, nucleic acids, small molecules and other biomolecular components, providing new opportunities for structure-guided drug discovery. AI can also assist in virtual screening and molecular optimization by prioritizing compounds with potentially favorable activity, selectivity, pharmacokinetic and toxicity profiles. However, challenges including data quality, model interpretability, algorithmic bias, chemical validity, reproducibility, computational requirements and experimental validation remain important limitations. Regulatory agencies have consequently begun developing frameworks for evaluating the credibility and appropriate use of AI-generated information in drug development and regulatory decision-making. This review discusses recent advances in AI-assisted drug design and discovery, with particular emphasis on machine learning, generative AI, molecular generation, structure prediction, virtual screening, drug repurposing, challenges and future perspectives.

.