Data from an Evaluation of ChatGPT for Nutrient Content Estimation from Meal Photographs

Background/Objectives: Advances in artificial intelligence now allow combined use of large language and vision models; however, there has been limited evaluation of their potential in dietary assessment. This data arose from a study that aimed to evaluate the accuracy of ChatGPT-4 in estimating nutritional content of commonly consumed meals from meal photographs.Methods: Meal photographs (n=114) were uploaded to ChatGPT, and it was asked to identify the foods in each meal, estimate their weight, and estimate the nutrient content of the meals for 16 nutrients for comparison with the known values. There were a total of 39 unique meals with each one photographed 3 times for 3 different portion sizes giving rise to 114 photographs. This dataset is in the form of an excel workbook containing four worksheets. The worksheet titled "ChatGPT Foods & Weights" contains the foods identified by ChatGPT in each of the 114 meal photographs as well as its estimate for the weight of each of those foods. The worksheet titled "Actual Foods & Weights" contains the true foods and weights for each of the meal photographs. The worksheet "ChatGPT Nutrition Estimates" contains ChatGPT's estimates of the nutrition content of each of the 114 meal photographs for 16 different nutrients. The worksheet "Actual Nutrition Content" contains the true nutrition content of the meals in the photographs.