Are Teens Getting Risky Diet Advice From AI Chatbots?

The rapid adoption of artificial intelligence tools has extended to health and wellness, with many teens turning to chatbots for information on everything from homework help to sensitive personal questions. But a new study raises concerns about the accuracy – and potential harm – of nutrition advice offered by these readily available platforms. Researchers have found that popular AI chatbots consistently generate meal plans for adolescents seeking weight loss that are nutritionally unbalanced, often falling short on essential carbohydrates while overemphasizing protein and fats.

The study, published March 12 in Frontiers in Nutrition, assessed the meal plans created by five widely used chatbots – ChatGPT-4o, Gemini 2.5 Pro, Claude 4.1, Bing Chat-5GPT and Perplexity – in response to prompts requesting three-day weight loss plans for hypothetical 15-year-olds. The prompts detailed the age, height, and weight of the teens, and asked for breakfast, lunch, dinner, and two snacks with portion sizes specified in grams or milliliters. Researchers, led by Betül Bilen of Istanbul Atlas University, found a consistent pattern: the AI-generated plans were significantly lower in calories and carbohydrates than those recommended by a registered dietitian for growing teenagers. Science News reports on the findings.

Why This Matters for Growing Adolescents

Adolescence is a period of rapid growth and development, requiring adequate nutrition for bone health, brain development, and overall well-being. Restrictive or unbalanced diets can interfere with these crucial processes. “Even if you are overweight, you still have that growth piece,” explains Taiya Bach, a registered dietician at the University of Wisconsin-Madison, as reported by Gizmodo. A significant portion of a teenager’s caloric intake supports hormonal changes and bone growth, needs that are not adequately addressed by the low-calorie, skewed meal plans generated by the chatbots.

On average, the AI-generated meal plans were approximately 695 calories per day below the dietitian-designed plans – equivalent to missing an entire meal. This caloric deficit, coupled with the macronutrient imbalance, raises concerns about potential negative health consequences.

The Study Details: How the AI Plans Were Evaluated

The research team crafted prompts in Turkish, then translated them into English for reporting purposes. They compared the chatbot-generated plans to dietitian-created plans for four hypothetical 15-year-olds – two classified as overweight and two as obese, with one male and one female in each category. The analysis revealed that while the chatbots varied in their specific recommendations, they consistently produced plans with lower carbohydrate levels and higher protein and fat content than recommended. The full study, published in Nutrients, details the methodology and findings.

Beyond Calorie Counts: The Risks of Poor Nutritional Guidance

The potential for harm extends beyond simply inadequate calorie intake. Stephanie Partridge, a public health and nutrition researcher at the University of Sydney, emphasizes that young people should not undertake restrictive eating without the supervision of healthcare professionals. A dietitian considers a range of individual factors – including underlying health conditions, socioeconomic status, and family dynamics – when developing a personalized diet plan. These nuances are unlikely to be accounted for by an AI chatbot.

Perhaps more concerning is the risk of fostering a harmful relationship with food. Restrictive diets, like those suggested by the chatbots, can increase the risk of developing disordered eating patterns. Weight loss, particularly for adolescents, is a sensitive issue, and relying on a non-specialized tool could exacerbate existing vulnerabilities.

How Widespread is This Practice?

According to the Pew Research Center, a substantial 64% of U.S. Teens report using AI chatbots. While the primary uses are information gathering and schoolwork assistance, the potential for teens to seek guidance on health and diet is significant. Although specific data on AI chatbot use for meal planning is currently limited, the increasing reliance on online resources for health information suggests that this is a growing trend.

Some dietitians, like Stephanie Kile with Equip, a virtual eating disorder treatment program, are already encountering patients who have been influenced by chatbot advice. Kile notes that patients sometimes prioritize the chatbot’s recommendations over her professional guidance, particularly when the AI reinforces pre-existing unhealthy beliefs about weight.

What’s Next: Addressing the Gap in Reliable Information

Researchers acknowledge that this study represents an initial step in understanding the potential risks of AI-generated nutrition advice. Further research is needed to examine how teens are actually using these tools in real-life scenarios and whether they are influencing eating behaviors. Betül Bilen and her colleagues emphasize the need for more investigation into AI usage patterns and the development of strategies to ensure that teens have access to accurate and reliable nutrition information.

Rebecca Raeside, a public health researcher at the University of Sydney, highlights the importance of involving teens in the research process and recognizing their awareness of the limitations of AI technology. She suggests that many young people view chatbots as a supplementary source of information rather than a definitive guide.

For parents and teens, the key takeaway is caution. While AI chatbots can be valuable tools for certain tasks, they should not be relied upon for personalized health or nutrition advice. Consulting with a registered dietitian or other qualified healthcare professional remains the safest and most effective approach to achieving healthy eating habits.