A Narrative Review of Integrating Artificial Intelligence and Omics in Food Science and Nutrition: Current Uses, Issues, and Future Perspectives.

This narrative review examines how artificial intelligence (AI) and omics are being integrated in food science and nutrition, with particular attention to current applications, practical limitations, and challenges that are specific to foods and dietary exposure. Foods are complex, multicomponent matrices whose composition and biological effects can change during processing, cooking, storage, and consumption. Dietary exposure also varies over time and across individuals. These features make food and nutrition fundamentally more difficult to model than many domains in which AI has advanced rapidly. Relevant literature was identified through searches of PubMed, Google Scholar, and Web of Science. For studies applying AI to foods, reports published in 2023 or later were preferentially selected unless suitable recent studies were unavailable. Selected publications were independently evaluated by the authors and subsequently discussed. The integration of AI and omics approaches are being applied to data integration, feature selection, predictive modeling, network analysis, clustering, interpretation, flavor and taste design, discovery of bioactive ingredients, functional food development, and prediction of dietary responses involving metabolites and the gut microbiome. Across these applications, food- and nutrition-specific difficulties emerged repeatedly, including the partial observability of food composition, changes caused by processing and cooking, mismatches between short-term measurements and long-term health outcomes, nonlinear dose-response relationships, and inter-individual variation. These challenges add to general issues such as data quality, multi-omics integration, model interpretability, and external validation. The integration of AI and omics is expanding the capacity to analyze complex data in food science and nutrition, but reliable translation will require methods adapted to the dynamic and partly observable nature of foods and diets. Improved standardization, exposure assessment, validation, study design, and interpretability will be essential for future research and practical application.