Redefining Preclinical Research Paradigms: AI-Driven Drug Discovery as a Transformative Approach to Accelerate Innovation, Improve Predictive Accuracy, and Reduce Reliance on Animal Testing

Drug discovery has historically been hindered by extended timelines, high costs, and low clinical success rates. Conventional methods such as high-throughput screening, structure-based design, and medicinal chemistry optimization, while scientifically valuable, often fail to deliver efficient translation into safe and effective therapeutics. Artificial intelligence (AI) and machine learning (ML) now provide unprecedented opportunities to accelerate every stage of drug discovery by leveraging large, heterogeneous datasets and powerful predictive algorithms. This manuscript presents a comprehensive review of AI-driven drug discovery, highlighting advances made between 2019 and 2024. Applications are critically examined across the pipeline: target identification, hit discovery, lead optimization, ADME-toxicity prediction, and clinical trial design. Special emphasis is given to transformative model architectures such as graph neural networks (GNNs), transformer models, and generative frameworks, as well as classical machine learning methods that remain relevant for specific tasks. Challenges including data quality, interpretability, regulatory acceptance, and ethical considerations are evaluated alongside strategies to mitigate bias and improve transparency. Case studies such as DiffDock for generative molecular docking, Trial Pathfinder for AI-based patient stratification, and Mol-BERT for chemical representation learning illustrate the tangible impact of these innovations. The manuscript concludes by identifying research gaps and future directions, including explainable AI (XAI), multimodal data integration, federated learning, and democratization of AI tools for global accessibility. Overall, AI is not simply a set of computational tools but a paradigm shift, offering a faster, more precise, and ethically responsible framework for pharmaceutical research and development. Keywords: Artificial intelligence (AI); Machine learning (ML); Drug discovery; Target identification; Lead optimization; Graph neural networks (GNNs); Transformer models; Generative AI; ADMET prediction; Clinical trial design; Drug repurposing; Ethical considerations; Explainable AI (XAI).