The researchers first screened 800 apps before selecting 18 for further evaluation. These 18 apps, which included both AI-powered nutrition apps and manual food logging apps, were then evaluated for their ability to recognize ingredients and estimate energy content.

Dr Juliana Chen, lead author of the study and a registered dietitian, lecturer and researcher in the discipline of nutrition and dietetics at the University of Sydney, suggests that while AI-embedded apps offer convenience over manual food recording, they should be used with caution.

“When patients or the public use apps to track their food intake or manage their weight, the process can often seem overwhelming,” Dr. Chen said. “Adding AI features like food image recognition could make the process much easier for everyone.”

“However, it is important to always check that the portion size detected matches what you have eaten. Some apps only identify foods, while others also estimate portion sizes and energy intake. So, for those who are losing weight, it is essential to check that the app’s estimates match what you have eaten.”

A key part of the study was to test the accuracy and adaptability of these apps to three different diets – Western, Asian and recommended (based on the Australian Dietary Guidelines) – to ensure a range of cultural food preferences were taken into account.

Under Dr Chen’s supervision, MSc nutrition and dietetics students Xinyi Li, Annabelle Yin and Ha Young Choi found that manual food recording apps overestimated energy intake from the Western diet by an average of 1,040 kilojoules, while they underestimated energy intake from the Asian diet and the recommended diet by an average of 1,520 kilojoules and 944 kilojoules respectively.

In contrast, AI-powered food apps often struggled to accurately identify the energy content of mixed Asian dishes. For example, beef pho overestimated calories by 49%, while pearl milk tea underestimated calories by up to 76%.

“AI-powered nutrition apps are generally better at detecting individual Western foods when they are separated on a plate,” said Dr. Chen, also of the Charles Perkins Center. “However, they often struggle with mixed dishes, such as spaghetti bolognese or burgers. This problem is more common with Asian dishes, which typically contain a variety of mixed components that may not be found in the respective app’s database, potentially leading to errors when calculating the energy content of a particular meal.”

The study recommends several steps to improve nutrition apps. These include ensuring that the educational content and advice provided by the apps is evidence-based and reliable, which can be achieved through collaboration with nutrition experts.

“To improve the credibility and accuracy of nutrition apps, creators should involve dietitians in their development, train AI models with diverse food images—especially for mixed and culturally diverse dishes—expand food composition databases, and educate users on capturing high-quality food images for better recognition accuracy,” Dr. Chen said.

“If you are monitoring your health, such as managing your high blood pressure or watching your sodium intake, it is important to compare your food choices to nutrition labels or consult a registered dietitian. The expertise of a dietitian is invaluable in these cases, as they can provide more accurate estimates of how much energy your body is consuming and what it needs most to achieve a healthy, holistic diet.”

This assessment was conducted using the Mobile Application Rating Scale (MARS) and the Application Behavior Change Scale (ABACUS).

Following the evaluation, “Noom” received an average score of 4.44 out of 5 on the MARS scale, meaning it scored highly in terms of engagement, functionality, aesthetics, and quality of information. It also received a perfect ABACUS score of 21/21 for integrating numerous features that promote behavior change, goal setting, tracking, and educational content.

Among other AI-based apps, “MyFitnessPal” and “Fastic” successfully recognized a sample of 22 images of various foods and drinks, achieving success rates of 97% and 92%, respectively.