The research on AI-powered calorie tracking apps reveals that their estimates often fall short of actual nutritional content. While these apps provide a convenient way for users to monitor their food intake using photos, studies indicate that they can underestimate calories and fat by about one-third.
In a recent study where meals were precisely measured, researchers tested four popular photo-based apps—MyFitnessPal, LoseIt!, CalAI, and Appediet—finding a significant discrepancy between the apps’ estimates and the actual nutritional content of the meals. On average, the apps missed 250 to 345 calories and underestimated fat content by approximately 30 grams. More reliable results were reported for higher-calorie meals but lower accuracy for lower-calorie meals and ketogenic dishes.
The findings suggest users should approach the results from these apps with caution, especially since they have been shown to underestimate fats. To enhance accuracy in calorie tracking, the researchers recommend complementing these tools with traditional dietary assessment methods.
