Related Experiment Video
Updated: May 21, 2025

Control of Eating Behavior Using a Novel Feedback System
Published on: May 8, 2018
Can artificial intelligence (AI) chatbot tools be used effectively for nutritional management in obesity?
Hatice Merve Bayram1, Zehra Margot Çelik2, Hatice Kübra Barcın Güzeldere3
1Department of Nutrition and Dietetics, Faculty of Health Science, Istanbul Gelisim University, Istanbul, Turkiye.
Artificial intelligence (AI) chatbots like ChatGPT-4o and ChatGPT-4 show limitations in providing accurate dietary plans for obesity, falling short on calorie and nutrient recommendations compared to Turkish Dietary Guidelines. Expert dietitians remain essential for clinical nutrition management.
Area of Science:
- Nutrition Science
- Artificial Intelligence
- Medical Nutrition Therapy
Background:
- Artificial intelligence (AI), specifically generative pre-trained transformer (ChatGPT) models, are increasingly explored for dietary planning in various health conditions.
- Previous research suggests AI's potential utility in assisting with nutritional management.
Purpose of the Study:
- To compare the accuracy and clarity of energy, macro, and micronutrient content in diet plans generated by ChatGPT-4o and ChatGPT-4 against the Turkish Dietary Guidelines (TDG)-2022 for obesity management.
- To evaluate the suitability of AI-generated menus for medical nutrition therapy.
Main Methods:
- A comparative content analysis was performed using ChatGPT-4o and ChatGPT-4 to create 1800 kcal daily diet plans for a 20-year-old female with obesity.
- AI models provided recommendations for dietary management, nutrition care processes, and menu planning.
- Outputs were evaluated by three registered dietitians and analyzed statistically.
Main Results:
- ChatGPT-generated menus were inconsistent with TDG-2022 recommendations, providing significantly lower calorie content than the "1800 kcal prompt" (P < 0.001 for ChatGPT-4o).
- Key nutrients, including fats (P = 0.003), carbohydrates (%), potassium, and calcium (P < 0.05), were inadequately represented compared to TDG-2022.
- Nutrient analysis indicated underperformance in meeting recommended intakes for critical micronutrients like calcium and an unbalanced macronutrient distribution.
Conclusions:
- ChatGPT-4o and ChatGPT-4 exhibit limitations in delivering precise and reliable dietary management plans for obesity.
- While AI chatbots can offer supplementary insights, they do not substitute the critical expertise of dietitians in clinical nutrition planning.
- Caution is advised when integrating AI tools into medical nutrition therapy due to current accuracy and completeness deficits.
Related Concept Videos
Obesity
Regulation of Food Intake
Hormonal Regulation
Role of Communication in the Nursing Process II: Planning and Implementation
Stress Prevention and Stress Management Techniques VI
Motivation and Self-Determination
Motivation, the driving force behind behavior, plays a pivotal role at every stage of the change process. The research...

