An NIH study found four popular photo-based calorie-tracking apps underestimated calories and fat by about a third.
Snap a photo, get a calorie count.
It sounds simple—but a new study finds some popular AI-powered food tracking apps may underestimate calories and fat by roughly a third.
Researchers at the National Institutes of Health tested four popular calorie-tracking apps using photos of more than 100 meals prepared in a tightly controlled metabolic kitchen.
Every ingredient was precisely weighed, allowing the apps' estimates to be compared with the meals' exact calories and nutrients.
The results: All four apps underestimated the calorie content of the meals by 250 to 345 calories, on average, and fat by about 30 grams.
The apps did a better job estimating carbohydrates than fat or protein.
One study author says, “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.”
A follow-up analysis of more than 200 additional meals also suggests the apps are less accurate with low-carb ketogenic meals,
likely because they consistently underestimate their higher fat content, according to the researchers.
They say combining AI photo analysis with more traditional food tracking methods could improve the apps' real-world accuracy.
Source: These findings were presented at Nutrition 2026 and should be considered preliminary until they are published in a peer-reviewed scientific journal.
Author Affiliations: National Institutes of Health
Copyright © 2026 HealthDay. All rights reserved.
Leave a Reply