In a test of four photo-based AI calorie-tracking apps, researchers found that calorie and fat estimates were about one-third too low on average. The study, presented by Olivia Charles at NUTRITION 2026, the flagship annual meeting of the American Society for Nutrition held July 25-28 in National Harbor, Maryland, used standardized photographs of 102 meals prepared in a controlled metabolic kitchen at the NIH Clinical Center. Researchers submitted the images to MyFitnessPal, LoseIt!, CalAI and Appediet. Across all four apps, estimated calorie totals were about 250 to 345 calories too low per meal on average, and the apps underestimated fat by approximately 30 grams. MyFitnessPal and LoseIt! were more accurate when analyzing higher-calorie meals than lower-calorie meals, and all four apps produced more consistent estimates for carbohydrates than for other macronutrients. Aaron Hengist, a postdoctoral visiting fellow with the Intramural Program of the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK), noted that the accuracy of many photo-based apps has not been thoroughly evaluated.