Toward AI-Driven Olfactory Biomarkers for Nutrition and Preventive Healthcare: A Literature Review

Artificial Intelligence (AI) is rapidly reshaping the sectors of digital health and preventive medicine, although the role of olfaction—one of the most informative yet underutilized human senses—remains largely unexplored in this context. A growing body of evidence links olfactory function to metabolic health, dietary habits, neurodegeneration, and overall well-being, suggesting that olfactory performance and Volatile Organic Compound (VOC) signatures, often characterized by making use of artificial olfactory tools or electronic noses, could serve as powerful, non-invasive disease biomarkers. Meanwhile, advances in machine learning, sensor technologies, and data-driven modeling open new avenues to transform olfactory inputs into actionable health indicators for early diagnosis, personalized nutrition, and sustainable healthcare practices. In such a framework, this narrative review synthesizes current knowledge on the relationship between olfaction, nutrition, and health, examines the emerging landscape of digital olfaction and AI-based olfactory biomarker analysis, and discusses how these technologies could enable the potential discovery of new biomarkers and the setup of clinical decision-support tools. We further highlight opportunities for AI-driven olfactory biomarkers to improve nutritional assessment, monitor dietary adherence, detect early metabolic or neurodegenerative changes, and promote novel food choices. Finally, we outline key challenges, such as data standardization, model explainability, sensor drift, privacy of biochemical signatures, and ethical implications, with the final aim to propose a roadmap for integrating olfactory data into trustworthy, secure, and scalable AI systems for preventive healthcare.