LO
CapabilityCognitive automation2024-08-05
Food-photo apps identified between 97% and 46% of food components, and their automatic energy estimates were inaccurate, a comparison of seven apps found
Dietitian / nutritionistoccupation page →
Event date / reported
2024-08-05
Evidence stage
CapabilityA demo, benchmark or paper shows the task can be done. Updates what the technology can do — not what employers will do.
Tasks this bears on
Assessing what someone eats
Recording and estimating a person's intake from food diaries, recalls and photos.
Being augmented≈ Platform inference
Where this applies
A study of nutrition apps, seven of which offered AI food-image recognition, tested on food images including mixed dishes. The most accurate identified 97% of food components (38 of 39) and the least accurate 46% (18 of 39); the authors found automatic energy estimations from AI food-image recognition inaccurate and say collaborating with dietitians is essential to improve the apps. Consumer apps on a small image set.
What this means
Recognising a food in a photo is solved better than knowing how much energy is on the plate.
What it does not yet show
Consumer apps on a small image set; not a test of dietitians' assessments.
What you can check
Open the Nutrients article on food logging and AI food image recognition and find "automatic energy estimations".
Does it change the assessment?
No — and this stage does not move it either. A "Capability" record is real evidence, but it does not upgrade a task judgement on its own. The 1 linked judgement above stand where they were.
Source
Evaluating the Quality and Comparative Validity of Manual Food Logging and Artificial Intelligence-Enabled Food Image Recognition in Apps for Nutrition Care, Nutrients 16(15), 2573 (2024; PMC11314244) · verified 2026-09-30 · Claude (VOLO agent) · interpreted 2026-09-30 · Claude (VOLO agent)
Primary source — published by the party that did this, or the authority of record. No co-signature needed.