AI Finds Simple Food Swaps That Make Meals Healthier and Cheaper

Key Takeaways:

• A new artificial intelligence framework can identify just one to three ingredient swaps that make a meal meaningfully more nutritious and noticeably less expensive, without overhauling it entirely.
  
• In testing, AI-generated meals sat 47% closer to United States Department of Agriculture (USDA) nutritional targets than the real meals they were modelled on, while staying close to people’s actual eating habits in type and flavour.
  
• Applying one to three swaps lifted nutritional quality by roughly 10% and cut modelled meal costs by 19 to 32%, most often by adding vegetables or legumes and removing high-sodium or processed items.
  
  
  

Turning nutrition science into practical meals

A new artificial intelligence framework that recommends as few as one to three ingredient swaps can make everyday meals meaningfully more nutritious and less expensive. That is the central finding of a study published on 28 May in the open-access journal PLOS Digital Health by PhD student Trevor Chan and Professor of Computer Science Ilias Tagkopoulos of the University of California, Davis.

The challenge the researchers set out to tackle is a familiar one. The dietary guidelines that help reduce people’s risk of conditions such as diabetes and cardiovascular disease are well established, yet turning that nutrition science into day-to-day meals remains difficult for most people. Many existing diet recommendation tools ask people to change too much all at once, which can lead to practices that are hard to sustain or leave people unsure how to put the advice into action.

How the study worked

To build their framework, the researchers drew on data from the What We Eat in America study, analysing 135,491 meals logged by 55,228 adults. From this, they identified common meal patterns across breakfast, lunch and dinner.

They then trained a generative AI model to create realistic meals that followed those patterns, while also adjusting serving sizes. Crucially, the team tested whether the model could pinpoint one, two or three ingredient swaps within each meal that would further improve both its nutritional value and its cost.

What the AI achieved

The results were striking. Compared with real meals in the same dietary pattern, the AI-generated meals were 47% closer to USDA nutritional targets, while remaining close in overall meal type and flavour to what people actually eat.

When ingredient substitutions were applied, swapping just one to three foods improved nutritional quality by approximately 10%, while reducing modelled meal costs by between 19 and 32%. The most common substitutions the system identified involved adding vegetables or legumes, and swapping out high-sodium or processed items.

Improving eating habits

The trained model also outperformed an unspecialised model. Set against GPT-4o, the purpose-built system produced meals that came closer to USDA guidelines on macronutrients.

The authors are careful to note the limits of the work. The evaluation is entirely computational and has not yet been tested with real users. Even so, they suggest the approach could help people find simple ways to improve their eating habits.

As the authors write: “By turning dietary guidelines into realistic, budget-aware meals and simple swaps, this framework can support public-health programs and consumer apps.”

Chan and Tagkopoulos summarise the thinking behind the study: “Dietary guidelines often tell people what a healthy diet should look like, but they do not always show how to get there from the meals people already eat. Our study shows that it is possible to translate dietary standards into practical meal-level changes by identifying a small number of ingredient substitutions that can make meals healthier and cost-effective, while keeping them recognizable…[w]hat we found most interesting is that improving meals does not necessarily require a complete redesign. In many cases, targeted substitutions may be enough to move a meal closer to dietary recommendations, which could make healthy eating feel more practical and achievable.”

They add: “Healthier eating does not have to mean giving up the meals people already enjoy. With AI, we can identify small ingredient substitutions that preserve taste, while are better for our health and our pocket.”

Funding

The work was supported by grants from the U.S. Department of Agriculture and the National Science Foundation.

Source: UC Davis College of Engineering