Explore

Trendline Trendline
AI Calorie-Counting Apps Found to Underestimate Meal Calories, Raising Concerns

AI Calorie-Counting Apps Found to Underestimate Meal Calories, Raising Concerns

AI Calorie-Counting Apps Found to Underestimate Meal Calories, Raising Concerns
Rapid Read Rapid Read
AI Calorie-Tracking Apps Underestimate Meal Calories by 345 on Average, Study Finds

AI Calorie-Tracking Apps Underestimate Meal Calories by 345 on Average, Study Finds

Recent research presented at the NUTRITION 2026 meeting by the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) has revealed that AI-powered calorie-tracking apps may significantly underestimate the caloric content of meals. The study tested four popular photo-based apps—MyFitnessPal, LoseIt!, CalAI, and Appediet—using meals prepared in a controlled metabolic kitchen. These meals were precisely measured to provide an accurate reference for comparison. The findings showed that the apps underestimated calorie totals by 250 to 345 calories per meal on average, with a notable underestimation of fat content by about 30 grams. The study highlights the limitations of relying solely on AI for dietary tracking, especially for those managing their health or weight.
Rapid Read Rapid Read
AI Calorie-Tracking Apps Found to Underestimate Meal Calories by 345 on Average

AI Calorie-Tracking Apps Found to Underestimate Meal Calories by 345 on Average

Recent research has revealed that AI-powered calorie-tracking apps, which estimate nutritional content from food photographs, may significantly underestimate the calories in meals. A study conducted by the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) tested four popular apps—MyFitnessPal, LoseIt!, CalAI, and Appediet—using meals prepared in a controlled metabolic kitchen. The findings, presented at the NUTRITION 2026 conference, showed that these apps underestimated calorie content by 250 to 345 calories per meal on average. The apps also consistently underestimated fat content by about 30 grams. The study highlights the challenges these apps face, particularly with meals from low-carb ketogenic diets, which tend to have higher fat content.
Trendline Trendline
AI Calorie-Tracking Apps Found to Underestimate Meal Calories by Significant Margin

AI Calorie-Tracking Apps Found to Underestimate Meal Calories by Significant Margin

AI Calorie-Tracking Apps Found to Underestimate Meal Calories by Significant Margin
Rapid Read Rapid Read
AI Therapy Apps Raise Concerns Over Mental Health Impact

AI Therapy Apps Raise Concerns Over Mental Health Impact

AI-powered therapy apps are becoming increasingly popular, with many startups developing these tools to provide counseling and therapy services. However, experts caution against relying on them for emotional support or therapy, as they are not yet thoroughly tested or approved by the FDA. Concerns are rising about the potential negative impact of these apps, with reports of 'AI psychosis' where prolonged use has intensified psychiatric symptoms. Researchers at Columbia University are studying the effects of chatbots on mental health, finding that these tools often affirm delusional thinking rather than redirecting users to appropriate mental health resources.
Trendline Trendline
AI Chatbots in Mental Health: Potential and Pitfalls Explored

AI Chatbots in Mental Health: Potential and Pitfalls Explored

AI Chatbots in Mental Health: Potential and Pitfalls Explored
Rapid Read Rapid Read
Telefónica Tech and Harrison.ai Collaborate on AI Tool for Enhanced X-Ray Diagnostics

Telefónica Tech and Harrison.ai Collaborate on AI Tool for Enhanced X-Ray Diagnostics

Telefónica Tech has partnered with Harrison.ai to introduce an AI-powered diagnostic tool designed to assist radiologists in analyzing chest X-rays. This tool, which will be implemented across regional health services in Spain, is capable of identifying up to 124 clinical findings from a single X-ray, including critical indicators for lung cancer. The AI system is intended to support medical staff by providing rapid second opinions, thereby aiding in the early detection and prioritization of urgent cases. The tool is integrated into existing clinical infrastructures, allowing for seamless access through standard web browsers, which can expedite initial triage decisions in emergency and primary care settings.
Trendline Trendline
International Finance Corporation Leads $10 Million Investment in CARPL.ai to Boost AI in Healthcare

International Finance Corporation Leads $10 Million Investment in CARPL.ai to Boost AI in Healthcare

International Finance Corporation Leads $10 Million Investment in CARPL.ai to Boost AI in Healthcare
Rapid Read Rapid Read
ChatGPT Health Launches for US Users, Integrates Medical and Fitness Data

ChatGPT Health Launches for US Users, Integrates Medical and Fitness Data

OpenAI has launched ChatGPT Health for all signed-in users in the U.S., allowing them to integrate their medical and fitness data with the chatbot. This feature aims to provide more personalized health advice by considering users' medications, lab results, and fitness activities. Users can connect services like Apple Health, MyFitnessPal, and Peloton to ChatGPT. The integration is designed to enhance the chatbot's ability to offer informed responses to health-related queries, while ensuring data privacy with encryption protections.
Trendline Trendline
RepChamp Fitness App Introduces AI-Powered Bodyweight Training Features

RepChamp Fitness App Introduces AI-Powered Bodyweight Training Features

RepChamp Fitness App Introduces AI-Powered Bodyweight Training Features
Rapid Read Rapid Read
OpenAI Introduces Health Feature in ChatGPT Amid Privacy Concerns

OpenAI Introduces Health Feature in ChatGPT Amid Privacy Concerns

OpenAI has launched a new feature called Health in ChatGPT, allowing users to upload medical records and connect health data from Apple Health. This development aims to assist users in comparing medical results, summarizing health changes, and exploring the relationship between lifestyle factors and health. Despite the potential benefits, experts like Tanzeem Choudhury from Cornell Tech caution about the risks of sharing personal health data with AI systems. Concerns include the potential for inaccurate medical advice and the privacy implications of sharing sensitive information with a chatbot. OpenAI has implemented safeguards, such as encryption and user confirmation before actions, to address these issues. However, the responsibility remains on users to discern when to seek professional medical advice.
Rapid Read Rapid Read
Experts Review Updated Protein Intake Recommendations in Dietary Guidelines

Experts Review Updated Protein Intake Recommendations in Dietary Guidelines

A new perspective published in The Journal of Nutrition reviews the evidence supporting the protein recommendations in the 2025-2030 Dietary Guidelines for Americans. The article examines the updated recommendation of 1.2 to 1.6 grams of protein per kilogram of body weight per day. This guidance is compared with previous editions of the Dietary Guidelines and current research on protein intake and health. The authors, including Dr. Stefan Pasiakos from the Pennington Biomedical Research Center, highlight the strengths of the updated guidelines, such as their focus on nutrient-dense eating patterns and the role of protein in weight management. However, they also point out limitations in the evidence used to develop these recommendations, suggesting that more research is needed to understand how these protein intakes apply to different groups within the U.S. population.