ChatGPT can now generate personalised meal plans in seconds, but experts warn it cannot replace a dietitian. A 2024 study showed chatbots struggled with complex health cases, often failing to provide accurate advice. While AI helps with daily planning, it lacks the deep medical insight needed for specific health conditions.
A diet plan tailored to your health needs is now just a prompt away. ChatGPT can generate a personalised seven-day meal plan in seconds. It tells you what to eat for breakfast, how much protein you need, which foods to avoid, and even what to cook for dinner. No appointment, no waiting, and no consultation fee. For anyone trying to make sense of nutrition, the appeal is obvious.
AI is rapidly becoming an informal nutrition adviser. People use chatbots to calculate calories, plan weight-loss diets, interpret food labels, identify nutrient deficiencies, and find alternatives for specific dietary needs. However, an important distinction exists between information that sounds personalised and advice that is nutritionally appropriate.
A 2024 study in the Journal of Clinical Medicine evaluated 10 AI chatbots using two hypothetical patient profiles. In the more complex case, involving a 65-year-old woman with type 2 diabetes, reduced muscle mass, and chronic kidney disease, no chatbot achieved 50% accuracy. Some even offered contradictory recommendations regarding protein intake. The researchers concluded that general-purpose chatbots could not replace an expert dietitian in complex clinical situations.
The problem becomes clearer when we consider what a diet must account for. Two people of the same age and weight can have very different nutritional requirements, depending on their medical history, activity levels, medications, digestive health and overall health. Someone managing diabetes, kidney disease or a hormonal condition cannot simply follow a generic high-protein or low-carbohydrate plan.
A 2025 study in Nutrients compared AI-generated weight-loss plans from three chatbots. While the plans scored reasonably well for overall diet quality, their macronutrient and fatty acid balance was consistently weak. In other words, a meal plan can look balanced on paper without getting every nutritional detail right. These studies tested specific chatbot versions and prompts, so their findings should not be taken as a verdict on all AI systems available today.
There is also the question of food culture. Nutrition advice is not simply about calories and protein. It needs to account for what people actually eat, their budgets, cooking habits, regional ingredients, and dietary preferences. A meal plan built around unfamiliar ingredients or requiring unrealistic preparation time is unlikely to work, no matter how impressive its nutritional calculations may be.
AI can help discover new ingredients, suggest recipe substitutions, and make everyday meal planning easier. But knowing which foods pair well together or how to assemble a balanced-looking meal is not the same as understanding what an individual's body actually needs. Nutrition is deeply personal, and a meal plan that works for one person may be entirely unsuitable for another.
The distinction matters. AI can be a useful starting point for understanding food and for making everyday meal planning easier. However, persistent symptoms, suspected deficiencies, supplement dosages, and diets for medical conditions require guidance from a qualified professional.
The smartest use of AI may not be to ask it what we should eat every day, but to use it to ask better questions about the food we eat. Convenience can make nutritional information more accessible, but it cannot replace the judgement that comes from understanding an individual, not merely their data.