Study Reveals AI Chatbots Give Teens Dangerous Diet Advice
AI Chatbots Give Teens Dangerous Diet Advice, Study Finds

A new study has revealed that artificial intelligence chatbots are providing teenagers with dangerously inadequate eating advice and diet plans that fail to meet basic nutritional requirements. Researchers from Istanbul Atlas University tested several popular AI models, including ChatGPT and Gemini, instructing them to create weight-loss meal plans for adolescents.

Alarming Calorie Deficits Found in AI-Generated Plans

The researchers discovered that, on average, the meal plans generated by AI contained nearly 700 fewer calories than those designed by qualified dietitians. This substantial deficit is equivalent to missing an entire meal each day, which could have serious clinical consequences for growing teenagers.

Specific Nutritional Imbalances Identified

Beyond the calorie shortfall, the study identified significant nutritional imbalances in the AI-generated plans. The chatbots recommended carbohydrate intake levels between 32 and 36 percent, far below the approximately 50 percent that many nutritionists consider appropriate for energy needs. Simultaneously, the AI suggestions included protein levels about 20 grams higher than those recommended by dietitians.

Wide Pickt banner — collaborative shopping lists app for Telegram, phone mockup with grocery list

Researchers Issue Strong Warnings

Dr. Ayşe Betül Bilen, the study's lead author, emphasized the particular risks during adolescence. "Following such unbalanced or overly restrictive meal plans during the teenage years may negatively affect growth, metabolic health, and eating behaviours," she stated. The research team examined plans from ChatGPT 4, Gemini 2.5 Pro, Bing Chat-5GPT, Claude 4.1, and Perplexity, comparing them against registered dietitian recommendations.

Methodology and Participant Profiles

The study involved creating three-day meal plans covering three meals and two snacks daily for four 15-year-olds with different body compositions. This included both overweight and obese teenagers of both genders. Researchers provided the chatbots with specific age, height, and weight information to generate personalized plans.

Critical Development Period at Risk

Dr. Bilen explained why these findings are particularly concerning: "Adolescence is a critical period for physical growth, bone development, and cognitive maturation. Lower energy and carbohydrate intake, combined with increased protein and fat ratios, may pose risks during this growth period." She noted that AI models appear trained to generate plausible-sounding responses rather than clinically precise nutritional guidance.

Expert Commentary on the Dangers

Anastasia Kalea, a senior nutrition researcher from University College London, provided additional context about the limitations of AI in this domain. "While digital platforms provide a stigma-free sanctuary for people living with obesity, the current reliance on AI tools presents a significant risk," she observed. Kalea highlighted that these models typically depend on generalized public data rather than the nuanced medical and dietary histories that healthcare professionals collect through careful assessment.

Calls for Safer AI Development

The research team has issued clear warnings to teenagers about using AI tools for diet planning and has called for the development of safer alternatives created with professional input. They emphasize that AI systems currently lack the personalization and clinical judgment that trained dietitians provide through active listening and comprehensive evaluation of individual circumstances.

Dr. Bilen summarized the core issue: "AI models are primarily trained to generate responses that appear plausible and user-friendly rather than clinically precise. Our findings suggest they may rely on generalized or popular diet patterns instead of fully integrating age-specific nutritional requirements." The Independent has contacted the AI companies mentioned in the study for their responses to these concerning findings.

Pickt after-article banner — collaborative shopping lists app with family illustration