Artificial intelligence is reshaping precision nutrition

Artificial intelligence (AI) is becoming one of the main drivers of precision nutrition, a field that aims to move beyond “one-size-fits-all” dietary advice and tailor nutritional recommendations to everyone’s characteristics.

The rationale is straightforward: people do not respond to the same food in the same way. For instance, variability in postprandial glycaemia, lipid metabolism, appetite, microbiome composition, physical activity, sleep, and clinical phenotype means that the same dietary pattern may be beneficial for one person and less effective for another.

AI is especially useful here because it can integrate large, diverse datasets (such as clinical, behavioural, biochemical, omics, and digital) to identify patterns that are difficult to detect with traditional statistical methods. Early work in this field showed that machine-learning models could predict individual postprandial glycaemic responses using clinical, dietary, lifestyle, and gut microbiome data, opening the way for data-driven personalised dietary advice [1-3].

Recent literature suggests that AI is now being utilised across at least four major domains of precision nutrition: dietary assessment, phenotyping and risk prediction, generation of personalised recommendations, and continuous feedback via digital platforms. A 2025 scoping review on AI for precision nutrition reported rapid growth in studies employing machine learning, deep learning, computer vision, and natural language processing, particularly in metabolic diseases and lifestyle-related conditions. Similarly, broader reviews published in 2025 describe AI as a key enabling technology for integrating dietary intake, biomarkers, continuous glucose monitoring, microbiome profiles, and patient-generated health data into more adaptable nutritional care models. In other words, AI does not replace nutritional science; it makes it more detailed, dynamic, and potentially more clinically actionable [4-7].

One of the clearest applications is dietary assessment, traditionally one of the weakest points in nutrition research and practice because it heavily relies on self-report. AI-based image recognition systems, wearable sensors, and language models can now assist in identifying foods, estimating portions, and converting meals into nutrient profiles with less manual effort. This is important because precision nutrition is only as reliable as the quality of the input data. A recent systematic review of AI-based dietary intake assessment found that these tools are promising and, in some settings, comparable to human estimation, although the evidence remains limited by small sample sizes, preclinical testing, and heterogeneity in validation methods. Therefore, AI may significantly improve data capture, but standardisation and external validation remain essential before routine large-scale implementation [8-10].

A second area is prediction: identifying who is likely to respond to a particular dietary intervention, and in which manner. This is where AI is particularly valuable, because nutritional responses are multifactorial. Recent research has continued to enhance predictive models by integrating continuous glucose monitoring, anthropometry, clinical chemistry, dietary records, and gut microbiome data. Some studies now also include serum metabolites and other multi-omics features to better stratify responses. Importantly, the field is shifting from proof-of-concept to conducting intervention studies. For instance, randomised trials and subsequent systematic reviews suggest that AI-supported personalised nutrition can lower postprandial glucose responses and may improve certain cardiometabolic outcomes, although results for weight loss and long-term behavioural changes are more modest and less consistent. The overall message from the literature is encouraging but cautious: precision nutrition appears biologically feasible and clinically promising, yet the strength of evidence varies across outcomes and populations [11-13].

The microbiome has become one of the most discussed elements of AI-driven precision nutrition. This is because gut microbial communities affect fibre fermentation, glycaemic response, bile acid metabolism, and inflammatory signalling, but their effects are highly individual. AI techniques are increasingly used to identify microbiome-derived signatures associated with responsiveness to specific diets or nutrients. Recent studies indicate that machine-learning models can classify individuals based on their potential benefit from fibre or other targeted interventions, and newer deep-learning methods seek to capture more complex diet-microbiome-host interactions. At the same time, this remains one of the most methodologically challenging fields, because microbiome data are noisy, context-dependent, and highly sensitive to variations in sequencing, preprocessing, and population characteristics [7, 14, 15].

Another important development is the expansion of precision nutrition beyond diabetes and obesity to include maternal and child health, gastrointestinal diseases, and critical care. A 2025 Nature Communications perspective highlighted the potential of AI and precision nutrition to enhance maternal and child health, especially in low-resource settings, where improved phenotyping and decision support could help allocate scarce resources more effectively. Other recent reviews discuss applications in IBS, ICU nutrition, and broader nutrition care pathways. This indicates that the relevance of AI in nutrition is expanding; it is no longer solely about “smart diets” for healthy adults but also about supporting clinical decision-making in more vulnerable populations [16-18].

However, the literature is equally clear about the limitations. AI systems in nutrition are often trained on small, demographically limited datasets or collected under highly controlled conditions. This raises concerns about bias, generalisability, and health equity. Additionally, many models remain difficult to interpret, which may reduce trust among clinicians and patients. Ethical analyses published in 2025 emphasise the need for transparency, explainability, privacy protection, and clear governance of data ownership, especially when recommendations are generated from sensitive biological and behavioural data. Generative AI adds further complexity: early systematic evidence suggests that large language model-based nutrition advice is promising, but its quality and reliability remain inconsistent enough that expert oversight continues to be essential [19-21].

In summary, the latest evidence indicates that AI can significantly enhance precision nutrition by improving dietary assessment, enabling multimodal phenotyping, and supporting more personalised and adaptive dietary recommendations. However, the field remains in a translational stage. The strongest message is not that AI has already solved personalised nutrition, but that it has provided the methodological framework to make it more scientifically feasible than ever before. The next step will be to demonstrate reproducible clinical benefits across diverse real-world populations. while ensuring that these technologies remain interpretable, equitable, and anchored to sound nutritional science.

References.

1. Zeevi D, et al. Personalized Nutrition by Prediction of Glycemic Responses. Cell. 2015;163(5):1079-1094. doi:10.1016/j.cell.2015.11.001
2. Berry SE, et al. Human postprandial responses to food and potential for precision nutrition. Nat Med. 2020;26(6):964-973. doi:10.1038/s41591-020-0934-0.
3. Ferreira DD, et al. Assessing the Links Between Artificial Intelligence and Precision Nutrition. Curr Nutr Rep. 2025;14(1):47. doi:10.1007/s13668-025-00635-2
4. Wu X, et al. A Scoping Review of Artificial Intelligence for Precision Nutrition. Adv Nutr. 2025;16(4):100398. doi: 10.1016/j.advnut.2025.100398.
5. Panayotova GG. Artificial Intelligence in Nutrition and Dietetics: A Comprehensive Review of Current Research. Healthcare (Basel). 2025;13(20):2579. doi:10.3390/healthcare13202579
6. Agrawal K, et al. Artificial intelligence in personalized nutrition and food manufacturing: a comprehensive review of methods, applications, and future directions. Front Nutr. 2025;12:1636980. doi:10.3389/fnut.2025.1636980
7. Mundt C, et al. AI-Driven Personalized Nutrition: Integrating Omics, Ethics, and Digital Health. Mol Nutr Food Res. 2025;69(24):e70293. doi: 10.1002/mnfr.70293.
8. Cofre S, et al. Validity and accuracy of artificial intelligence-based dietary intake assessment methods: a systematic review. Br J Nutr. 2025;133(9):1241-1253. doi:10.1017/S0007114525000522
9. Phalle A, et al. Navigating next-gen nutrition care using artificial intelligence-assisted dietary assessment tools-a scoping review of potential applications. Front Nutr. 2025;12:1518466. doi:10.3389/fnut.2025.1518466
10. Choochaiwattana W, et al. AI powered dietary proportion assessment for improving accuracy and practicality of the balanced meal plate model. Sci Rep 2025;15:45273. doi: 10.1038/s41598-025-29631-w
11. Rein M, et al. Effects of personalized diets by prediction of glycemic responses on glycemic control and metabolic health in newly diagnosed T2DM: a randomized dietary intervention pilot trial. BMC Med. 2022;20(1):56. doi: 10.1186/s12916-022-02254-y.
12. Munawaroh EF, et al. Effectiveness of Personalized Nutrition on Management Diabetes Mellitus Type 2 and Prediabetes in Adults: A Systematic Review. Diabetes Metab Syndr Obes. 2025;18:2783-2796. doi: 10.2147/DMSO.S528619.
13. Bermingham KM, et al. Effects of a personalized nutrition program on cardiometabolic health: a randomized controlled trial. Nat Med. 2024;30(7):1888-1897. doi: 10.1038/s41591-024-02951-6.
14. Song D, et al. Gut microbiome predicts personalized responses to dietary fiber in prediabetes: a randomized, open-label trial. Nat Commun. 2025;16(1):11506. doi: 10.1038/s41467-025-66498-x.
15. Liu YY. Deep learning for microbiome-informed precision nutrition. Natl Sci Rev. 2025;12(6):nwaf148. doi: 10.1093/nsr/nwaf148.
16. Mehta S, et al. Advances in artificial intelligence and precision nutrition approaches to improve maternal and child health in low resource settings. Nat Commun. 2025;16(1):7673.
17. Wang X, et al. Artificial Intelligence Applications to Personalized Dietary Recommendations: A Systematic Review. Healthcare (Basel). 2025;13(12):1417. doi:10.3390/healthcare13121417
18. Briassoulis G, et al. AI-Enabled Precision Nutrition in the ICU: A Narrative Review and Implementation Roadmap. Nutrients. 2025;18(1):110. doi:10.3390/nu18010110
19. Abrahams M, et al. Perspective on the ethics of AI at the intersection of nutrition and behaviour change. Front Aging. 2025;6:1423759. doi: 10.3389/fragi.2025.1423759.
20. Brankovic A, et al. Perspectives, challenges and future of artificial intelligence in personalised nutrition research. Proc Nutr Soc. 2025. doi:10.1017/S0029665125100657
21. Hieronimus B, et al.  GenAI in nutritional sciences (GAINS): A systematic review and reporting framework for future research. Nutr Res. 2025;143:66-77. doi:10.1016/j.nutres.2025.09.011