Improving Nutrition Information Quality with Google AI 

Nearly three-quarters of people in the U.S. turn to the internet first for information about health and wellness, and they frequently ask questions about nutrition. However, online nutrition information is often conflicting and sometimes contains misinformation. New technology, including artificial intelligence (AI), represents an exciting new opportunity to consistently present reliable access to high-quality nutrition information to more people. However, the nutrition and health evidence that current AI tools use to provide information to consumers lacks nutritional nuance, consensus, and accuracy. Ongoing basic standards for elevating particular trusted sources are not keeping up with the breadth, pace, and highly variable quality of nutrition information available nor the growing consumer demand. Importantly, AI-based tools cannot solve this dilemma without strategic, effective curation of inputs and testing and validation of outputs. This project, Improving Nutrition Information Quality with Google AI, aims to solve these issues by developing a trusted AI-based tool to deliver the most up-to-date nutritional advice and facts, driven by the expertise of the research team, including multiple registered dietitians.  

Aim 1

Produce a first-in-class AI-based tool to elevate trusted nutrition information and mitigate nutrition misinformation, powered by Palm2 and Tufts FIMI nutrition expertise. 

Aim 2

Generate a conversational response between the end-user and Google AI model to resemble a registered dietitian visit.

Aim 3

Consider how Google AI can generate personalized guidance for different end users and subsequent platforms, including members of the public, healthcare payers and providers, online food retailers, employer wellness plans, life insurance plans, and more.