An artificial intelligence framework that suggests swapping just one to three ingredients can make meals meaningfully more nutritious and cheaper, according to a new study published in an open access journal. PLOS Digital Health By Trevor Chan and Elias Tagkopoulos of the University of California, Davis, USA.
Although dietary guidelines to reduce the risk of diseases such as diabetes and cardiovascular disease are well established, incorporating nutrition into the daily diet remains difficult for most people. Many diet recommendation tools ask people to make too many changes at once, leading to unsustainable practices and confusion about how to implement the changes.
In the new study, researchers used data from 135,491 meals recorded by 55,228 adults in the What We Eat in America survey to identify common dietary patterns for breakfast, lunch, and dinner. It then trained a generative AI model to follow those patterns to create realistic meals while adjusting serving sizes. The researchers then tested whether the AI could identify one, two, or three ingredient swaps in each meal to further improve nutrition and cost.
When compared to real meals with the same dietary pattern, the AI-generated meals came 47% closer to USDA nutritional goals, while the overall meal variety and taste remained close to what humans actually eat. When applying ingredient substitution, exchanging one to three foods increased nutritional quality by approximately 10% and reduced modeled meal costs by 22% to 34%. The most common substitutions identified by the system included adding vegetables and legumes and replacing high-salt and processed foods.
Compared to the unspecialized model GPT-4o, the trained model produced a diet closer to USDA guidelines for macronutrients. The authors emphasize that this evaluation is entirely computational and has not been tested with real users. But they suggest it could help people identify simple ways to improve their eating habits.
”By turning dietary guidelines into realistic, budget-conscious meals and easy swaps, this framework can support public health programs and consumer apps.” writes the author.
Chan and Tagkopoulos summarize: “Dietary guidelines often tell people what a healthy diet should look like, but they don’t always tell people how to get there from the diet they already eat. Our research aims to transform dietary standards into actual diet-level changes by identifying substitutions for a small number of ingredients that can make meals healthier and more cost-effective. “What we found most interesting is that improving your diet doesn’t necessarily require a complete redesign. In some cases, just taking targeted substitutions can bring your diet closer to the recommendations, making healthy eating more realistic and achievable.”
“They added.”Eating healthier doesn’t mean giving up the foods people already enjoy. AI can be used to identify small ingredient substitutions that are healthier and wallet-friendly while preserving taste.”
sauce:
Reference magazines:
Chan, T. others. (2026) Translate dietary standards into healthy diets with few substitute ingredients. PLOS Digital Health. DOI: 10.1371/journal.pdig.0001367. https://journals.plos.org/digitalhealth/article?id=10.1371/journal.pdig.0001367
