Abstract:

Artificial intelligence (AI) is transforming the landscape of drug design and discovery by expediting key steps and improving success rates. This review provides a comprehensive overview of how AI techniques are applied across the drug development pipeline, from early target identification through clinical trial optimisation. Traditional drug discovery is notoriously time- consuming (often over a decade) and costly (over \$1 billion) with a high failure rate of >90% (Dermawan & Alotaiq, 2025; Ocana et al., 2025). In contrast, AI-driven approaches leverage big data and machine learning to identify novel drug targets, virtually screen large chemical libraries, design new drug molecules de novo, and predict crucial properties like ADMET (absorption, distribution, metabolism, excretion, toxicity) with greater speed and accuracy. We discuss sub- fields including AI-based target identification (e.g. using network analysis and knowledge graphs), virtual screening and molecular docking, generative models for de novo drug design, ADMET and toxicity prediction, and AI enhancements in clinical trial design (such as patient selection and adaptive trials). Notably, several AI-designed drug candidates have progressed to clinical trials in recent years, with an observed Phase I success rate around 80–90%, significantly higher than historical averages of ~40–60% (Sale et al., 2025). We highlight case studies such as the AI-guided repurposing of baricitinib for COVID-19 and the discovery of novel molecules (e.g. a DDR1 kinase inhibitor identified in 21 days). Key challenges—data quality, model interpretability, and regulatory acceptance—are also examined. Overall, AI has demonstrated the potential to accelerate drug discovery timelines, reduce costs, and yield innovative therapeutics, particularly in areas like oncology, infectious disease, and neurology. Continued advances and careful integration of AI with human expertise will be critical to fully realise its benefits in developing safer, more effective drugs.

Keywords:

Artificial Intelligence, Drug Discovery, Machine Learning, De Novo Drug Design, Virtual Screening, Target Identification, ADMET Prediction, Clinical Trial Optimisation.

Aim/Objectives

Methods