To read this content please select one of the options below:

AI showdown: info accuracy on protein quality content in foods from ChatGPT 3.5, ChatGPT 4, bard AI and bing chat

Hati̇ce Merve Bayram (Department of Nutrition and Dietetics, Istanbul Gelisim University, Istanbul, Turkey)
Arda Ozturkcan (Department of Nutrition and Dietetics, Istanbul Gelisim University, Istanbul, Turkey)

British Food Journal

ISSN: 0007-070X

Article publication date: 9 July 2024

Issue publication date: 15 August 2024

251

Abstract

Purpose

This study aims to assess the effectiveness of different AI models in accurately aggregating information about the protein quality (PQ) content of food items using four artificial intelligence (AI) models -– ChatGPT 3.5, ChatGPT 4, Bard AI and Bing Chat.

Design/methodology/approach

A total of 22 food items, curated from the Food and Agriculture Organisation (FAO) of the United Nations (UN) report, were input into each model. These items were characterised by their PQ content according to the Digestible Indispensable Amino Acid Score (DIAAS).

Findings

Bing Chat was the most accurate AI assistant with a mean accuracy rate of 63.6% for all analyses, followed by ChatGPT 4 with 60.6%. ChatGPT 4 (Cohen’s kappa: 0.718, p < 0.001) and ChatGPT 3.5 (Cohen’s kappa: 0.636, p: 0.002) showed substantial agreement between baseline and 2nd analysis, whereas they showed a moderate agreement between baseline and 3rd analysis (Cohen’s kappa: 0.538, p: 0.011 for ChatGPT 4 and Cohen’s kappa: 0.455, p: 0.030 for ChatGPT 3.5).

Originality/value

This study provides an initial insight into how emerging AI models assess and classify nutrient content pertinent to nutritional knowledge. Further research into the real-world implementation of AI for nutritional advice is essential as the technology develops.

Keywords

Acknowledgements

The manuscript presents a research study that used AI chatbots for its investigations. In particular, it involved the use of ChatGPT versions GPT-3.5 and GPT-4.0, both developed by OpenAI. Additionally, this study used Bing Chat and Bard AI. All authors had access to the data and played a substantial role in writing the manuscript.

Citation

Bayram, H.M. and Ozturkcan, A. (2024), "AI showdown: info accuracy on protein quality content in foods from ChatGPT 3.5, ChatGPT 4, bard AI and bing chat", British Food Journal, Vol. 126 No. 9, pp. 3335-3346. https://doi.org/10.1108/BFJ-02-2024-0158

Publisher

:

Emerald Publishing Limited

Copyright © 2024, Emerald Publishing Limited

Related articles