(P03-049-24) The Role of Artificial Intelligence in Nutritional Dietary Assessment: An Article Review

Objectives: The purpose of this article review was to investigate the role of artificial intelligence (AI) in nutrition research, specifically in dietary assessment.

Methods: Eight electronic databases were searched: PubMed, Web of Science, EBSCO, Agricola, Scopus, IEEE Explore, Google Scholar, and Cochrane for studies identification. Articles underwent different screening phases: elimination of duplicates, title and abstract review and full-text review. This review followed the PRISMA guidelines for articles selection from 2000-2023. A total of 1737 articles were retrieved, of which 22 were included for analysis and their quality was determined.

Results: The focus of most articles is on AI’s ability to operate as a nutritionist. Some of the constrains on data collection in dietary assessment are accuracy and time. AI has proven to have the capacity to evaluate diets, while removing biases and the burden of self-reported data. However, AI in nutrition is still not very common and is in the early development stages. Thus, AI is still limited in food recognition and food estimation. Ethics is one of the main concerns about the use of AI. Interestingly, this issue is not being entirely addressed and remains unsolved.

Conclusions: AI in the nutrition field still needs more research. Nonetheless, it is gaining more importance in this field. Health improvement using AI, AI validation, AI accuracy and patients’ satisfaction in nutrition using AI are yet to be determined. Therefore, clinical research is needed to determine the efficacy of AI in dietary assessment. Ethical guidelines to avoid collateral damage to certain populations, as a result of AI models being trained by humans also needs to be addressed.

Funding Sources: No funding source