Applications of Machine Learning in Assessing Human Dietary Exposure to Environmental Chemical Contaminants

Diet is a major pathway for human exposure to environmental contaminants. Researchers have established associations between dietary intake and increased body burden of environmental chemicals. Traditional human dietary exposure assessment has limitations in accuracy, high-throughput, predictive ability, and ability to explore complex interrelations. With recent advances in artificial intelligence, this narrative review examined the applications of machine learning (ML) in human dietary exposure assessment. We conducted a search using terms/synonyms related to machine learning, environmental chemicals, diet, and exposure. Study criteria were established, and two reviewers conducted screening from February-April 2025. Nineteen (19) studies published from 2019-2025 were included. Key applications identified were detecting/quantifying contaminants in food, predicting contaminant levels in the body, developing exposure profiles, and assessing health risks from interactions among diet, contaminants, and epidemiological factors. Studies utilizing ML demonstrated improved accuracy and predictive ability compared to traditional approaches. They also provided faster, more cost-effective, and less invasive animal testing for food safety. Challenges identified include poor model performance in certain instances, limited independent external validation, limited applicability, and limited complete datasets. Our review highlights ML as a valuable tool for human dietary exposure assessment and provides key future directions in this field.