Nutritional adequacy of AI-generated menus for institutional foodservice systems across different life stages: A theoretical menu analysis

Background Generative artificial intelligence (AI) systems are increasingly being explored for menu planning in institutional foodservice settings; however, evidence regarding the nutritional adequacy of AI-generated menus remains limited. Aim This study aimed to evaluate and compare the nutritional adequacy of menus generated by ChatGPT, Claude, and Gemini for three institutional foodservice settings serving different life stages: kindergarten, university cafeteria, and nursing home. Methods A theoretical menu analysis was conducted using 4-week weekday menu cycles generated by ChatGPT, Claude, and Gemini. Menus were developed for three institutional settings, yielding 180 daily menus. Energy, nutrient, and food group contents were analyzed using the Nutrition Information System (BEBIS 9.0), and adequacy was assessed according to the Turkish Dietary Guidelines. Summary Exploratory comparisons revealed differences in the nutritional composition of the generated menu sets across institutional settings. Mean energy content ranged from 887.9 to 1161.0 kcal in kindergarten menus and from 842.9 to 1189.0 kcal in university lunch menus (p < 0.001). Claude-generated menus generally contained higher amounts of protein and several micronutrients, whereas Gemini-generated menus provided lower amounts of certain nutrients, including folate, vitamin C, potassium, and calcium. Excessive estimated sodium content was a consistent concern across all institutional settings, while calcium contributions were generally low in kindergarten and university menus. These findings highlight the importance of professional nutritional evaluation of AI-generated menus, particularly with respect to sodium content and micronutrient adequacy. Given the single-generation design, the observed differences should be interpreted as exploratory findings specific to the evaluated menu cycles.

Authors

Publication Details

Journal
Nutrition and Health
Published
2026-11-19
DOI
https://doi.org/10.1177/02601060261508812
Primary Topic
Dietetics, Nutrition, and Education
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Nutritional adequacy of AI-generated menus for institutional foodservice systems across different life stages: A theoretical menu analysis

Bahar Yalçın
Nutrition and Health
Dietetics, Nutrition, and Education
article

Nutritional adequacy of AI-generated menus for institutional foodservice systems across different life stages: A theoretical menu analysis

Bahar Yalçın
article en

Abstract

Background Generative artificial intelligence (AI) systems are increasingly being explored for menu planning in institutional foodservice settings; however, evidence regarding the nutritional adequacy of AI-generated menus remains limited. Aim This study aimed to evaluate and compare the nutritional adequacy of menus generated by ChatGPT, Claude, and Gemini for three institutional foodservice settings serving different life stages: kindergarten, university cafeteria, and nursing home. Methods A theoretical menu analysis was conducted using 4-week weekday menu cycles generated by ChatGPT, Claude, and Gemini. Menus were developed for three institutional settings, yielding 180 daily menus. Energy, nutrient, and food group contents were analyzed using the Nutrition Information System (BEBIS 9.0), and adequacy was assessed according to the Turkish Dietary Guidelines. Summary Exploratory comparisons revealed differences in the nutritional composition of the generated menu sets across institutional settings. Mean energy content ranged from 887.9 to 1161.0 kcal in kindergarten menus and from 842.9 to 1189.0 kcal in university lunch menus (p < 0.001). Claude-generated menus generally contained higher amounts of protein and several micronutrients, whereas Gemini-generated menus provided lower amounts of certain nutrients, including folate, vitamin C, potassium, and calcium. Excessive estimated sodium content was a consistent concern across all institutional settings, while calcium contributions were generally low in kindergarten and university menus. These findings highlight the importance of professional nutritional evaluation of AI-generated menus, particularly with respect to sodium content and micronutrient adequacy. Given the single-generation design, the observed differences should be interpreted as exploratory findings specific to the evaluated menu cycles.

Nutrition and Health
Dietetics, Nutrition, and Education
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon