A university student prepares for an all-nighter to cram for finals. Wanting to stay alert without overdoing the caffeine, he needs help choosing an energy drink that supports mental focus and avoids a sugar crash. Meanwhile, a basketball player gets ready to play an intense three-hour game. She is looking for a low sugar drink that will replenish essential electrolytes and contains enough carbohydrates to keep up her energy levels throughout the game. Which beverage should she choose?
These are two real-world scenarios from Healthy Choice, an innovative technology-based simulation platform designed to help people develop nutrition literacy skills, or the ability to understand and use nutrition information to make healthier decisions. Healthy Choice was designed by Shan Li, a learning scientist and educational technologist at the College of Health. He uses advanced technologies to promote learning and instruction, and uncover behavior or cognitive processes in problem solving. He is an assistant professor, department of community and global health within the College of Health, as well as an affiliated faculty member in the Teaching, Learning and Technology program in the College of Education.
Li notes that around 50% of American adults struggle with health information, and limited nutrition literacy leads to poor dietary choices. To help solve this problem, learners assume the role of a health professional in Healthy Choice, recommending healthier food options for virtual customers. They review a real-world scenario to understand their customer’s needs and preferences, and are then presented with a selection of products commonly found in the grocery store. Learners carefully evaluate the nutrition information, ingredient lists, and marketing claims of each option and make an initial assessment of whether it meets the needs of the customer.
They select beverage options and rank them based on their relative importance. Throughout the process, an integrated ChatGPT function allows them to ask questions, clarify concepts, and receive guidance. When they submit their final choice, they are asked to write a summary reflecting on their decision-making process and to provide a justification for their choice. They then receive feedback indicating the correct answer and reasoning behind that beverage.
“Through these experiences, they develop 21st century skills, such as critical thinking, problem-solving, and self-reflection,” said Li. “They gain nutrition literacy. This is not just a platform for nutrition education, but they need to analyze the different kinds of items and use their critical thinking skills to choose the best option for their virtual customer.”
According to Li, Healthy Choice represents a shift in health education, transitioning from passive information delivery to active skill development. The platform’s design draws upon four major theoretical foundations, including situated learning theory, in which learning is embedded in authentic, real-world contexts; and scaffolding theory, in which AI provides adaptive support tailored to individual needs.
In a manuscript entitled “A Theory-Driven and AI-Enhanced Simulation Platform for Cultivating Nutrition Literacy,” Shan details his mixed methods study of Healthy Choice. 114 participants from Lehigh University evaluated the Healthy Choice platform and gave it high marks, rating it 8.5 out of 10 for ease of use. They appreciated the authenticity of the scenarios, as well as the interactive learning, support and clarification from AI, and structured decision-making tools. The study showed that AI can play a valuable role in health education and support integrating technology in course offerings.
In a separate mixed methods study detailed in “The Impact of Learner Characteristics and AI Utilization on Learning Experiences,” Li explored how individual differences influenced AI-supported learning in nutrition education. He found that participants with higher baseline nutrition literacy were less likely to rely on AI, instead trusting their own expertise. According to Li, this may be explained by the Expertise Reversal Effect, which suggests that instructional support that benefits students with limited prior knowledge may be unnecessary or even detrimental to more knowledgeable students.
“Prior knowledge is a stronger predictor of metacognitive judgments than AI assistance,” he said. “Simply providing AI access doesn’t automatically enhance learning confidence in this process.”
Li also found that individuals showed a strong preference for using AI for simple, factual information rather than for more complex analysis, and that to effectively integrate AI into education, the learner’s needs, the nature of the task, and the type of support provided must all be aligned. In the future we need to develop more adaptive AI system design to better support their learning, he said.
While currently at a pilot stage at the university level, the ultimate goal for Healthy Choice is to provide the general public with a platform for enhancing their nutrition literacy for leading a healthier lifestyle. The platform can be geared for younger students in grades K-12 as well.
