AI calorie-tracking apps miss up to 345 calories per meal, NIH study reveals

AI Calorie-Tracking Apps Face Accuracy Crisis

AI calorie-tracking apps that promise effortless nutrition monitoring through simple food photos are significantly underestimating what users actually consume, according to new research from the National Institutes of Health. In a rigorous evaluation of four popular AI-powered photo-based calorie-tracking apps, researchers discovered calorie underestimation ranging from 250 to 345 calories per meal on average, with fat content missed by approximately 30 grams1

1

. The findings were presented at NUTRITION 2026, the flagship annual meeting of the American Society for Nutrition, held July 25-28 in National Harbor, Maryland2

2

.

Aaron Hengist, a postdoctoral visiting fellow with the Intramural Program of the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK), emphasized the significance of these results. "Photo-based calorie tracking apps are very popular, especially for people trying to manage their health or lose weight," Hengist noted. "However, the accuracy of many of these apps has not been thoroughly evaluated. Our study helps address this question by looking at whether these apps can reliably estimate calories"1

1

.

Testing Photo-Based Calorie Tracking Against Precise Standards

The study leveraged meals prepared in a controlled metabolic kitchen at the NIH Clinical Center, where ingredients are measured to the nearest 0.1 gram. This meticulous preparation provided researchers with an exceptionally accurate reference point for evaluating app performance. Olivia Charles, a postbaccalaureate intramural research training fellow at NIDDK who presented the findings, and her team collected standardized photographs of 102 meals originally prepared for a broader nutrition study investigating how the body processes nutrients on either a low-carbohydrate ketogenic diet or a standard diet1

1

.

These images were submitted to MyFitnessPal, LoseIt!, CalAI and Appediet to determine how closely each app's estimates matched the known nutritional content. "By using meals prepared in a tightly controlled metabolic kitchen, we were able to compare the apps' estimates against a precise reference," said Hengist. "This kind of direct, high-quality comparison hasn't been available before"2

2

.

Understanding the Limitations of AI-Based Calorie Tracking

Photo-based calorie tracking relies on AI image recognition to identify foods shown in a picture and estimate portion sizes. The app then compares those estimates with nutrition databases to calculate calories and other nutrients1

1

. However, the study revealed significant inconsistencies. MyFitnessPal and LoseIt! demonstrated better accuracy when analyzing higher-calorie meals compared to lower-calorie ones, while all four apps produced more consistent estimates for carbohydrates than for other macronutrients1

1

.

Preliminary findings from testing more than 200 additional meals suggest that apps struggle particularly with ketogenic diets. These meals often contain more fat, which the apps tended to underestimate consistently1

1

. This pattern matters significantly for users following specific dietary protocols where macronutrient ratios are critical.

What This Means for Health-Conscious Users

The research carries immediate implications for millions who rely on these tools for weight management and health monitoring. "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," Hengist advised. "These apps tend to underestimate calories, especially from fats, so what they actually ate is likely higher than what the app shows"2

2

.

For users attempting to create a calorie deficit for weight loss, this underestimation could explain unexpected plateaus or slower-than-anticipated progress. Someone believing they've consumed 1,500 calories might have actually eaten closer to 1,800 or more. The researchers suggest combining photo-based tools with traditional methods of evaluating food intake and diet quality could make calorie tracking more accurate in everyday use1

1

. While these findings are preliminary and have not yet completed full peer review for publication in a scientific journal, they highlight a critical gap between the convenience of AI-powered nutritional estimation and the precision required for effective health management.🟡 compliments and suggestions=