Artificial Intelligence in Preformulation and Physicochemical Property Prediction: A Comprehensive Review

Artificial intelligence (AI) has gained significant attention in pharmaceutical sciences due to its ability to analyze complex datasets and generate reliable predictions. In preformulation studies, understanding the physicochemical properties of active pharmaceutical ingredients (APIs) is crucial for the development of safe, effective, and stable dosage forms. Conventional experimental approaches, although well established, are often time-consuming and resource intensive. AI-based techniques such as machine learning, deep learning, and quantitative structure–property relationship (QSPR) modeling provide efficient alternatives by enabling early prediction of solubility, stability, permeability, polymorphism, pKa, and lipophilicity. This review elaborates on the principles of AI, its applications in preformulation, commonly used algorithms, data sources, advantages, limitations, regulatory considerations, and future prospects, with emphasis on its relevance in modern drug development.