The future of artificial intelligence-driven personalized nutrition in gastroenterology and hepatology: Emerging trends and perspectives

Nutrition is central to preventing and managing gastrointestinal (GI) and liver diseases, yet dietary advice remains generic. Conditions such as IBD, IBS, MASLD, celiac disease, and GERD are strongly diet‑influenced. Conventional personalized nutrition fails to account for metabolic, genetic, and microbiome variability. AI integrates genomics, microbiome, metabolic markers, and real‑time dietary tracking to generate individualized recommendations. AI tools are advancing dietary assessment, condition‑specific optimization, and continuous monitoring via wearables and natural language processing. Challenges remain: lack of standardized protocols, bias, privacy concerns, limited clinical validation, and underrepresentation of nutrition in AI applications. Opportunities include federated learning, real‑world validation, and regulatory frameworks. This narrative review synthesizes literature from 2015–2025.

Introduction

AI, including machine learning and deep learning, has transformed GI and liver care—enhancing polyp detection, fibrosis staging, and outcome prediction. Nutrition, however, remains poorly integrated. Conventional dietary guidelines ignore individual variation in metabolism, genetics, microbiota, and lifestyle. Personalized nutrition aims to tailor advice; AI provides the computational scale by integrating multi‑omics and wearable data. In IBS, ML models identify dietary triggers; in MASLD, AI predicts responses to calorie‑restricted and macronutrient‑modified diets. Large language models (LLMs) such as ChatGPT enable real‑time, multilingual nutrition counseling, while mobile apps support symptom tracking. Yet nutrition‑specific predictive modeling is limited, and clinical validation, algorithmic transparency, and equity concerns persist.

Data Source

A narrative search of PubMed, Scopus, Web of Science, ScienceDirect, and Google Scholar (2015–2025) using keywords including AI, machine learning, deep learning, LLMs, personalized nutrition, gastroenterology, hepatology, IBD, IBS, MASLD, GERD, and celiac disease. Additional references were identified through citation tracking. This is not a systematic review.

Applications

AI spans screening/triage, automation of food‑diary analysis and symptom monitoring, and real‑time decision support. Generative AI (GPT‑4, LLaMA‑3) can deliver culturally tailored dietary guidance and behavior‑change techniques, but risks hallucinations and unsafe advice in complex patients; supervised fine‑tuning and retrieval‑augmented generation are being tested. Wearables, mobile platforms, and EHR‑integrated systems enable continuous monitoring and dynamic adjustments. AI apps such as Heali improve low‑FODMAP adherence in IBS. Customized LLMs (e.g., GERD Tool for Surgery) outperform general GPT‑4 in guideline alignment.

Microbiome‑based personalization: In IBS, an AI‑generated microbiota‑guided diet (ENBIOSIS) reduced symptom severity more than low‑FODMAP diets and improved microbial diversity. Another trial achieved 91% classification accuracy (AUC 0.964) for microbiome‑based IBS phenotypes. A multicenter RCT confirmed AI‑assisted diets outperformed FODMAP across all IBS subtypes. ML models using microbiome, metabolomics, and breath volatile organic compounds classify IBS variants with >90% accuracy in many studies. In IBD, AI‑integrated mobile tools enable early flare‑up identification and tailored guidance.

MASLD and hepatology: A study of >800 individuals and 46,898 meals used continuous glucose monitoring, clinical, behavioral, and microbiome data to predict postprandial glycemic responses; AI‑generated diets improved glycemic control and microbial composition. Digital phenotyping via smart scales, sleep monitors, and activity trackers supports adaptive adjustments. Integrative omics incorporating PNPLA3, TM6SF2, microbial taxa, and diet generate targeted prescriptions reducing steatosis. AI also assists with fibrosis staging, hepatotoxicity prediction, and post‑transplant risk stratification.

Challenges and Future Perspectives

Ethics and transparency: Deep learning and generative models are often “black boxes,” complicating consent and trust. Genomic and microbiome data raise privacy and discrimination concerns requiring legislative oversight.

Technical limitations: Data heterogeneity from food diaries, wearables, and sequencing introduces bias; standardization of formats and pipelines is essential. Federated learning can preserve privacy but requires infrastructure. Most studies are retrospective and single‑center; large prospective trials are needed.

Bias and equity: Datasets often underrepresent diverse populations; sociocultural dietary patterns and digital inequality risk inequitable recommendations. Financial and infrastructure barriers may exclude under‑resourced settings.

Acceptance and education: Clinicians need workflow integration and preserved decision‑making roles; patients may worry about impersonality and privacy. Interdisciplinary training across nutrition, computer science, and medicine is essential.

Opportunities: Synthetic datasets, platforms like AI4FoodDB and ENBIOSIS, and causal inference from longitudinal data. Regulatory oversight should treat AI as medical devices, ensuring transparency, validation thresholds, and liability clarity, alongside privacy safeguards and equitable access.

Conclusions

AI can transform personalized nutrition in gastroenterology and hepatology by integrating clinical, behavioral, and omics data into adaptive dietary strategies. Applications in IBS, IBD, and MASLD show improved individualization and outcomes. However, this narrative review lacks systematic rigor, and many applications remain preliminary. Bridging gaps requires interdisciplinary collaboration, robust validation, data standardization, and digital equity. Successful integration can make digestive and liver care more precise, proactive, and patient‑centered.

Full text

https://www.xiahepublishing.com/2994-8754/JTG-2025-00043

The study was recently published in the Journal of Translational Gastroenterology.

Journal of Translational Gastroenterology (JTG) dedicates to improving clinical diagnosis and treatment, advancing understanding of the molecular mechanisms, and promoting translation from bench to bedside of gastrointestinal, hepatobiliary, and pancreatic diseases. The aim of JTG is to provide a forum for the exchange of ideas and concepts on basic, translational, and clinical aspects of gastroenterology, and promote cross-disciplinary research and collaboration.

Follow us on X: @xiahepublishing

Follow us on LinkedIn: Xia & He Publishing Inc.