Artificial Intelligence in Drug Discovery and Development: Transforming Target Identification, Molecular Design, and Clinical Trial Prediction
Authors
Oliver T. Harrington
Francis Crick Institute and Department of Structural Biology, University College London, London WC2A 2AE, United Kingdom
Author
Mei-Lin Zhang
Vector Institute for Artificial Intelligence and Department of Pharmaceutical Sciences, University of Toronto, Toronto, ON M5G 1M1, Canada
Author
Fiona C. O'Brien
Francis Crick Institute and Department of Structural Biology, University College London, London WC2A 2AE, United Kingdom
Author
Keywords:
Artificial Intelligence, Drug Discovery, Deep Learning, Alphabet, Generative Chemistry, ADMET Prediction, Drug Repurposing, Deepmind, Vector Institute, Machine Learning, Precision Medicine
Abstract
The pharmaceutical drug discovery pipeline is characterised by high attrition rates, costs averaging USD 2.6 billion per approved drug, and timelines spanning 12-15 years. Artificial intelligence (AI) and machine learning (ML) technologies are fundamentally reshaping this landscape. The United Kingdom-home to DeepMind (London), BenevolentAI (London), and Esscientia (Oxford)- and Canada-home to the Vector Institute (Toronto) and the Montreal AI Ethics Institute- together constitute the most productive non-US AI drug discovery ecosystem globally. This review comprehensively examines the application of AI methodologies including deep learning, graph neural networks, reinforcement learning, and generative AI to drug target identification, de novo molecular design, ADMET property prediction, drug repurposing, and patient stratification. Landmark applications reviewed include AlphaFold2-informed target biology, generative chemistry platforms, and AI-driven biomarker discovery. We critically evaluate limitations including dataset biases, interpretability challenges, and validation gaps, and discuss the evolving regulatory frameworks at the MHRA (UK) and Health Canada. The review concludes by outlining the most promising near-term opportunities and critical unresolved challenges.
