Millions of people rely on AI-powered calorie tracking apps to count what they eat, snapping a quick photo of their meal and trusting the algorithm to do the math. But new research from the National Institutes of Health suggests the numbers on screen are dangerously low, potentially concealing hundreds of calories from the foods you think you know.
Scientists tested four popular photo-based apps: MyFitnessPal, LoseIt!, CalAI, and Appediet. Using meals prepared in a tightly controlled metabolic kitchen at the NIH Clinical Center, where ingredients are measured to the nearest tenth of a gram, researchers compared what the apps estimated against what was actually there. The gap was substantial.
Across all four apps, calorie estimates fell short by an average of 250 to 345 calories per meal. The apps also underestimated fat content by roughly 30 grams. In other words, if an app told you that your lunch contained 600 calories, you were likely consuming closer to 900.
The researchers collected standardized photographs of 102 meals from an ongoing diet study and submitted them to each app. MyFitnessPal and LoseIt! performed somewhat better with higher-calorie meals but struggled with smaller portions. All four apps were more reliable at estimating carbohydrates than fats or calories overall.
Aaron Hengist, a postdoctoral researcher at the National Institute of Diabetes and Digestive and Kidney Diseases, acknowledged the appeal of the technology. "Photo-based calorie tracking apps are very popular, especially for people trying to manage their health or lose weight," he said. "However, the accuracy of many of these apps has not been thoroughly evaluated."
The problem becomes more pronounced with certain types of meals. Preliminary analysis of more than 200 additional meals suggests that ketogenic and low-carb diets, which tend to contain more fat, pose particular challenges for the apps. The algorithms consistently underestimated fat content in these meals, leading to larger errors overall.
Hengist warned users to be skeptical of what appears on their screens. "People using a photo-based tracking app without adjusting the portions or entering the amounts of food should take the results with a grain of salt," he said. "These apps tend to underestimate calories, especially from fats, so what they actually ate is likely higher than what the app shows."
The research, presented at the American Society for Nutrition's annual meeting in July, suggests a practical solution: combining photo-based tracking with traditional food intake methods could improve accuracy. That might mean manually verifying portions or keeping a written log alongside the app.
For anyone relying on these tools to hit calorie or macronutrient goals, the findings carry real consequences. Someone attempting to maintain a 2,000-calorie diet could unknowingly consume an extra 1,000 calories per day if they trust their app across three meals. Over a week, that's 7,000 extra calories, or roughly two pounds of weight gain.
Author Jessica Williams: "These apps are convenient, but they're not honest about their own limitations, and users paying for them deserve better transparency about the margin of error."
Comments