DOI: 10.38079/igusabder.1803108 ISSN: 2536-4499

Artificial Intelligence in Nutrition & Dietetics (2010–2025): A Global Bibliometric and Evidence Mapping Study

Sedat Arslan
Show PDF Cite
Aim: This study aimed to map global research activity on artificial intelligence (AI) in nutrition and dietetics between 2010 and 2025, identify major thematic areas, and determine methodological and translational gaps that influence clinical application.Method: We used a descriptive bibliometric and evidence-mapping design, reported in line with PRISMA-ScR. Records published between 2010 and 2025 were retrieved from Web of Science (Core Collection), Scopus, and PubMed using AI- and nutrition-related keywords (full strategies in Supplement S1). After harmonization and two-stage deduplication (DOI → normalized title+year), the final corpus comprised 2.853 unique records spanning 2011–2024. We summarized descriptive indicators (annual production, journals, countries) and constructed keyword co-occurrence networks. Thematic clusters and methodological characteristics—data types, AI modalities, validation practices, and reporting signals—were evaluated against TRIPOD+AI, CONSORT-AI, SPIRIT-AI, and DECIDE-AI frameworks. No human participants or identifiable data were involved; ethics approval was not required.Results: The number of publications increased markedly after 2019. Three dominant themes were identified: (i) clinical risk prediction and decision support using tabular and electronic health record data; (ii) image-based dietary assessment and nutrient-intake estimation; and (iii) personalization and counseling enabled by large language models (LLMs). Although performance metrics were frequently reported, external validation, calibration, fairness, and openness of code/data remained inconsistent across studies.Conclusion: Research on AI in nutrition and dietetics has grown rapidly in the past decade, centering on predictive modeling, image-based assessment, and LLM-assisted personalization. However, methodological rigor and real-world validation are still limited. Future studies should prioritize standardized reporting, external validation, fairness assessment, and open science practices to ensure safe and generalizable implementation of AI in nutrition care.

More from our Archive