Abstract Cancer remains one of the leading causes of death worldwide and continues to pose a serious public health challenge. The limited success of many current treatments—often due to toxicity, poor selectivity, and the development of drug resistance—highlights the need for new and more effective therapeutic options. Phytochemicals have emerged as a valuable source of anticancer agents, offering rich structural diversity and a wide range of biological activities. However, identifying promising compounds from the vast chemical space of natural products remains difficult using conventional screening methods, which are typically slow, costly, and inefficient. In recent years, artificial intelligence (AI) has begun to transform phytochemical-based drug discovery. Machine learning and deep learning approaches are now used to support key steps in the discovery process, including metabolite identification, virtual screening, target prediction, and toxicity assessment. By integrating chemical, biological, and multi-omics data, AI enables a more systematic and data-driven exploration of natural product diversity. Despite these advances, challenges persist, particularly the scarcity of high-quality experimental data, the structural complexity of phytochemicals, and their limited representation in public databases. This review critically examines current AI applications in phytochemical-based anticancer drug discovery and discusses emerging strategies aimed at overcoming these limitations. Overall, AI-driven phytochemical screening represents a promising path toward accelerating the development of next-generation cancer therapies. Graphical Abstract
Santiago et al. (Tue,) studied this question.