Researchers in Sri Lanka tested a new smartphone app called FRANI that uses artificial intelligence to identify and track food eaten by teenage girls. They compared this AI app to traditional methods of measuring what people eat, like writing down food or recalling meals from memory. The AI app worked surprisingly well—it was just as accurate as the traditional methods but much easier to use. This is important because many developing countries struggle to afford expensive ways to measure nutrition, and this app could help doctors and health workers understand what teenagers are actually eating and how to help them stay healthy.
The Quick Take
- What they studied: Can a smartphone app powered by artificial intelligence accurately measure what food teenage girls eat compared to traditional methods?
- Who participated: 60 teenage girls aged 14-18 years living in cities and towns in Sri Lanka. The study was conducted in urban and semi-urban communities.
- Key finding: The AI app (FRANI) was just as accurate as traditional food tracking methods. It made very few mistakes when identifying foods—only 2% of foods were missed, compared to 12% missed by the traditional recall method.
- What it means for you: If you’re a teenager or parent in a developing country, this app could become a useful, affordable tool to track nutrition. However, this study was done in Sri Lanka, so results may vary in other countries. Talk to a healthcare provider before relying on any app for medical decisions.
The Research Details
Researchers recruited 60 teenage girls in Sri Lanka and asked them to record everything they ate over two separate days using three different methods. The first method was the FRANI app—a smartphone application that uses AI to recognize foods from photos. The second method was writing down and weighing every single food item eaten (considered the most accurate traditional method). The third method was asking girls to remember everything they ate in the past 24 hours (a common but less accurate method). By comparing all three methods, researchers could see how well the app worked compared to the gold-standard weighed food records and the traditional recall method that many doctors use.
This research approach is important because it tests the app in real-world conditions with actual teenagers in a developing country. Many new nutrition apps are only tested in wealthy countries with different foods and eating patterns. By validating the app against the most accurate method (weighed food records), researchers can prove it actually works before recommending it for widespread use. This helps ensure the app is reliable before it’s used to make health decisions.
This study has several strengths: it compared the app to the most accurate method available (weighed food records), it tested it in a real-world setting with actual teenagers, and it measured specific nutrients rather than just calories. However, the sample size was relatively small (60 girls), and the study was only done in one country, so results may not apply everywhere. The study was published in a respected nutrition journal, which suggests it met high scientific standards.
What the Results Show
The FRANI app performed very well when compared to weighed food records. For energy (calories) and vitamin A, the app was accurate within 10%—meaning it measured these nutrients almost perfectly. For protein, fiber, iron, and zinc, it was accurate within 15%. For fat, niacin, and folate, it was accurate within 20%. These results show the app can reliably measure most nutrients that teenagers consume. The agreement between the app and weighed food records ranged from moderate to very strong (scores of 0.49 to 0.89 on a scale where 1.0 is perfect agreement). Most importantly, the app made very few mistakes—it missed only 2% of foods that were actually eaten, which is much better than the traditional 24-hour recall method that missed 12% of foods.
When researchers compared the traditional 24-hour recall method (where people remember what they ate) to weighed food records, they found it was less accurate than the FRANI app. No nutrients fell within the 10% accuracy range for the recall method. This suggests that asking people to remember their meals is harder than using an app that identifies foods from photos. The FRANI app also made fewer mistakes overall—it incorrectly added foods that weren’t eaten only 7% of the time, compared to 9% for the recall method.
This research builds on growing evidence that technology-assisted dietary assessment tools can work well in developing countries. Previous studies have shown that traditional methods like food recalls are expensive and prone to errors, especially in countries with limited resources. This study shows that AI-powered apps may be a better solution. The FRANI app performed better than the traditional 24-hour recall method, suggesting it could replace older methods in similar settings. However, more research is needed in other countries and with different age groups to confirm these findings apply broadly.
The study had several limitations worth noting. First, only 60 teenage girls participated, which is a relatively small group. Second, the study was only conducted in Sri Lanka, so the results may not apply to teenagers in other countries with different foods and eating habits. Third, the study only looked at girls aged 14-18, so we don’t know if the app works as well for younger children or boys. Fourth, the study was conducted over just two days, so we don’t know if the app remains accurate over longer periods. Finally, the app requires smartphones and internet access, which may not be available to all teenagers in developing countries.
The Bottom Line
The FRANI app appears to be a reliable tool for measuring what teenagers eat in developing countries like Sri Lanka. It may be particularly useful in settings where traditional dietary assessment methods are too expensive or difficult to use. However, this recommendation is based on one study in one country, so more research is needed. Healthcare providers should consider this app as a potential tool, but it should not replace professional medical advice about nutrition. (Confidence level: Moderate—based on a single validation study with a small sample size)
This research is most relevant to: (1) teenagers and their parents in developing countries who want to track nutrition, (2) healthcare workers and nutritionists in low- and middle-income countries looking for affordable assessment tools, (3) public health programs trying to understand adolescent nutrition in developing regions, and (4) technology developers creating health apps for resource-limited settings. This research is less relevant to people in wealthy countries with access to traditional nutrition services, though the app may still be useful for convenience.
If you start using the FRANI app to track your diet, you should see accurate nutritional information immediately—the app provides feedback right after you photograph your food. However, to see actual health benefits from improving your diet based on the app’s information, you would typically need to make changes consistently for several weeks to months. Visible changes in energy levels or weight might take 4-8 weeks of consistent dietary changes.
Want to Apply This Research?
- Use the FRANI app to photograph and log meals for 3 days per week, tracking specific nutrients like protein, iron, and calcium intake. Set a weekly goal (for example, ‘meet iron recommendations 4 out of 7 days’) and review progress every Sunday.
- Start by using the app to photograph one meal per day for one week to get comfortable with it. Once you’re familiar with how it works, gradually increase to logging all meals. Use the app’s nutrient feedback to identify which nutrients you’re getting enough of and which ones you need more of, then make small changes (like adding more beans for iron or dairy for calcium).
- Track your nutrient intake weekly using the app’s summary features. Every two weeks, review which nutrients are consistently low and set one specific goal to improve (for example, ‘add one iron-rich food daily’). After one month, assess whether the changes you made are sustainable and adjust your approach if needed.
This research describes a validation study of the FRANI app in Sri Lanka and should not be considered medical advice. The app’s accuracy may vary in different countries, age groups, and settings. Before making significant dietary changes based on any app’s recommendations, consult with a healthcare provider, registered dietitian, or nutritionist. This app is a tool to help track food intake but should not replace professional medical guidance. If you have specific health conditions, allergies, or dietary restrictions, discuss app use with your healthcare provider. The study was conducted with a small sample size in one country, so individual results may vary.
This research translation is published by Gram Research, the science division of Gram, an AI-powered nutrition tracking app.