Development and Evaluation of an AI Program for Generating Personalized Meal Plans with a Perceived Quality Scale

• Grant number: 25/06258-7
• Support Opportunities: Scholarships in Brazil - Master
• Start date: January 01, 2026
• End date: February 28, 2027
• Field of knowledge: Health Sciences - Nutrition
• Principal Investigator: Fabiana Braga Benatti
• Grantee: Ellis Wollis Malta Abhulime
• Host Institution: Faculdade de Ciências Aplicadas (FCA). Universidade Estadual de Campinas (UNICAMP). Limeira , SP, Brazil

Abstract

This project aims to develop and evaluate a computer program leveraging Artificial Intelligence (AI) to generate personalized meal plans and assess their perceived quality by users. The research is structured in three main phases: (1) Prototyping of the AI system using advanced Large Language Models (LLMs) via API and prompt engineering techniques, integrating data from Brazilian food composition tables (TACO, TBCA) and nutritional guidelines (Brazilian Dietary Guidelines, WHO, DRIs, etc.); (2) Blind validation of the AI-generated meal plans by expert nutritionists, who will technically evaluate the recommendations without knowing their origin, comparing them to standard plans; and (3) testing with end-users, measuring usability (including System Usability Scale - SUS) and satisfaction with the generated plans. The project will develop a modular architecture system capable of producing balanced and tailored menus, ensuring nutritional safety and high acceptance. Innovative aspects include iterative evaluation through expert and user feedback (iterative feedback), and strict ethical compliance (informed consent with right of withdrawal, data privacy per LGPD, and Ethics Committee approval). Expected results include evidence of the technical quality and perceived quality of AI-generated meal plans, contributing to advances at the intersection of nutrition and technology, and enhancing the scalability of diet personalization.