AI-generated dietary advice risks teens’ nutrient intake and may trigger eating disorders

A study has found that teens using AI models for meal plans and dietary advice may be eating too few calories compared to those following plans created by dieticians. While undercalculating macronutrients such as carbs, the tools overcalculated proteins and lipids. 

This poses potential health hazards for adolescents at a crucial stage of growth and development if their diet lacks essential nutrients. Meanwhile, it may also trigger unhealthy eating behaviors for teenagers with body dissatisfaction.

“We show that diet plans generated by AI models tend to substantially underestimate total energy and key nutrient intake when compared to guideline-based plans prepared by a dietitian,” said Dr. Ayşe Betül Bilen, an assistant professor at the Faculty of Health Sciences at Istanbul Atlas University, Turkey.

“Following such unbalanced or overly restrictive meal plans during the teenage years may negatively affect growth, metabolic health, and eating behaviors.”

Under and overestimating

The study, published in Frontiers in Nutrition, used the AI models ChatGPT-4o, Gemini 2.5 Pro, Bing Chat-5GPT, Claude 4.1, and Perplexity. These tools can generate personalized diet plans, and all have free versions that are more accessible and commonly used by adolescent populations.

The researchers used prompts to create three meals and two snacks per day.The researchers used prompts to create three meals and two snacks per day, including the person’s height, weight, and age.

The tools created meal plans for four 15-year-olds. These were tailored for one boy and one girl classified as overweight with a body mass index (BMI) of around 25, and one boy and one girl classified as obese with BMI of around 30.

The AI models suggested diets with 700 calories less than those made by dieticians specializing in adolescent diseases. The authors stress that this difference is equivalent to a full meal and could cause clinical consequences.

The AI models also recommended a significantly lower daily energy intake from carbohydrates of 32–36%.

Meanwhile, it overcalculated crucial macronutrients. The AI tools recommended 20 g more protein per day than the dieticians’ recommendations, equating to 21–24% of daily energy intake coming from proteins. It also recommended that 41–45% of daily energy intake should come from lipids, which is also “way higher than dieticians’ recommendations,” assert the authors.

In comparison, the dieticians recommended daily energy intake consisting of protein 15–20%, lipids 30–35%, and carbohydrates 45–50%.

Problematic for adolescence

Many teens are turning to AI for weight-loss advice. Bilen says that AI-generated diet plans consistently deviate from the recommended macronutrient balance, which is particularly problematic for adolescents. This is due to the rapid growth and critical development period.

AI nutrition advice may influence unhealthy eating behavior, the researchers say.Following these AI recommendations and consuming 700 calories below the daily limit could jeopardize these developmental needs.

Additionally, it may also influence unhealthy eating behavior, such as post-meal vomiting or consuming laxatives. The researchers cite prior studies showing that 18.1% of students reported body dissatisfaction, 24.8% reported a desire to be thinner, and 10.5% reported unhealthy weight-control behaviors.

“Taken together, these findings suggest that a notable proportion of adolescents with pronounced body image and weight concerns rely on internet- and AI-based sources for nutrition information, and that the quality and safety of information obtained from these tools may be clinically important,” reads the study.

The authors say that AI-guided guidelines may be based on general or popular diet patterns rather than on integrating age-specific nutritional requirements.

They caution teenagers using AI tools to be cautious and to avoid restrictive diets or those based on excess protein or fat.

This issue stretches beyond adolescents, as a recent US national survey found that nearly half of its citizens rely on unaccredited sources, social media, and AI-generated recommendations for nutrition advice rather than trained professionals. It flags that consumers struggle to differentiate reliable data from misinformation.