Voice of Urdu – International News in English & Urdu

AI Calorie-Tracking Apps Underestimate Intake by 345 Calories, NIH Study Finds — AI-generated representative image. A smartphone displays a calorie-tracking app while held over a meal, illustrating the NIH study that found AI photo-based app

Editorial Team

NIH researchers found photo-based nutrition apps misjudge fat by 30 grams per meal, with ketogenic diets posing the greatest AI recognition challenge.

Popular AI-powered photo calorie-tracking apps underestimate caloric intake by as much as 345 calories per meal on average, according to new research from the National Institutes of Health. The study, which tested MyFitnessPal, LoseIt!, CalAI and Appediet against 102 precisely measured meals prepared in a controlled metabolic kitchen, also found that fat content was underestimated by approximately 30 grams per meal.

The findings raise significant concerns for millions of users who rely on photo-based tracking apps to manage their weight or monitor health conditions. If users consistently receive estimates that are roughly one-third too low, they may unknowingly consume far more calories and fat than intended, potentially undermining weight loss efforts and dietary management.

Key Research Findings

Across all four apps tested, calorie estimates fell short by a range of 250 to 345 calories per meal. Fat content was consistently underestimated by roughly 30 grams. Carbohydrate estimates proved more reliable than those for other macronutrients across all apps.

MyFitnessPal and LoseIt! demonstrated greater accuracy when analyzing higher-calorie meals compared to lower-calorie ones. In a follow-up analysis of more than 200 additional meals, preliminary results indicated that ketogenic diet meals, which typically contain higher fat content, posed greater challenges for the AI systems.

How Photo-Based Calorie Tracking Works

Photo-based calorie tracking apps use AI image recognition to identify foods in a photograph and estimate portion sizes. Those estimates are then cross-referenced with nutrition databases to calculate calorie counts and macronutrient breakdowns. Despite their widespread popularity, the accuracy of these tools has not been rigorously evaluated until now.

The study was conducted as part of a broader nutrition research project at the NIH Clinical Center examining how the body processes nutrients on low-carbohydrate ketogenic diets versus standard diets. Meals in the clinical trial are prepared in a metabolic kitchen where ingredients are measured to the nearest 0.1 gram, giving researchers an unusually precise reference point for evaluating the apps.

Researchers' Assessment

"By using meals prepared in a tightly controlled metabolic kitchen, we were able to compare the apps' estimates against a precise reference," said Aaron Hengist, a postdoctoral visiting fellow with the Intramural Program of the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK). "This kind of direct, high-quality comparison hasn't been available before."

Hengist cautioned that users relying solely on photo-based tracking without manually adjusting portions or entering food amounts should interpret results carefully. "These apps tend to underestimate calories, especially from fats, so what they actually ate is likely higher than what the app shows," he said. The research team suggests combining photo-based tools with traditional dietary assessment methods to improve everyday accuracy.

Preliminary Status and Next Steps

The findings were presented by Olivia Charles, a postbaccalaureate intramural research training fellow at NIDDK, on July 25 during the President's Oral Session at NUTRITION 2026, the American Society for Nutrition's flagship annual meeting held in National Harbor, Maryland. The abstracts have been reviewed and selected by an expert committee but have not yet completed full peer review. The results remain preliminary until published in a peer-reviewed scientific journal.

Quantum Tunneling Discovery Boosts Triplet Energy Transfer at Room Temperature

Chinese researchers unveil a proton shuttle mechanism that harnesses quantum effects to accelerate triplet energy migration, opening new paths for solar cells, lasers, and catalysis.