AI in Traditional and Herbal Medicine

Abstract

This review surveys the convergence of artificial intelligence (AI), machine learning (ML), natural language processing (NLP), and data science (DS) with traditional and herbal medicine on a global scale. We synthesize data landscapes, methodological advances, and translational pathways that enable rigorous, reproducible, and ethically responsible integration of AI-driven insights with traditional knowledge. Core topics include data sources spanning herbal constituent databases, ethnobotanical records, pharmacology and omics datasets, and analytical chemistry repositories; AI/ML/NLP/DS methodologies such as graph neural networks, multi-omics integration, QSAR, feature selection, NLP for literature-based knowledge discovery, computer vision for species identification and quality assessment, and spectroscopy-driven learning for quality control and adulteration detection; and application domains including diagnosis support, formula optimization, herb–drug interaction prediction, safety pharmacovigilance, and quality assurance. We emphasize interpretability, reproducibility, and data provenance, and we devote substantial attention to regulatory, ethical, and socio-economic dimensions, including benefit-sharing, traditional knowledge protection, and post-market surveillance of herbal products. Challenges such as data heterogeneity, standardization, and cross-cultural validation are analyzed, and a practical research roadmap is proposed to accelerate safe, evidence-based AI-enabled integration of traditional and herbal medicines into modern healthcare. The review identifies critical gaps and outlines concrete paths for future work, including standardized benchmarks, open datasets, and robust pipelines that can be adopted by researchers pursuing doctoral-level contributions.