Previously, obtaining a nutritional plan required visiting a nutritionist and sitting down to explain daily habits, medical history, and goals for losing weight or improving health. Today, with busy schedules of work, appointments, and attempts to commit to sports, artificial intelligence is able, within seconds, to suggest complete nutritional programs, calculate calories, provide meal replacements, and even monitor the user’s daily progress.
These robots seem like “nutritionists on the phone,” offering quick and affordable solutions to sometimes complex problems, gradually blurring the line between general advice and therapeutic nutritional advice. As reliance on these digital tools increases, a fundamental question arises: Can the nutritional advice provided by artificial intelligence be trusted, or does our health still need the human eye and professional expertise?
Read also
list of 2 itemsend of list
Doctor not yet certified
Recent studies suggest that AI is capable of providing good initial nutritional advice, but has great difficulty taking into account medical history, lifestyle, and individual needs. AI is already a powerful tool, but it does not replace the expertise of a dietitian who has the ability to clinically examine and understand human context.
In the end, the nutritionist does not only give general recommendations, but rather analyzes several factors before suggesting a specific regimen: medical history, medications, work pattern, movement, sleep, degree of stress, eating habits, and whether the person smokes or not. As for artificial intelligence, it does not evaluate your personal condition in this depth.
If you ask a robot to create a plan for you to lose 3 kilograms per week, it will take the request as is and suggest a system that achieves the goal. Although healthy weight loss usually ranges between 0.5 and 1 kilogram per week, he often just mentions this in a sideline or a note at the end of the conversation. As for the doctor and the nutritionist, they have an ethical responsibility and a degree of compassion that makes the way they deal with the request completely different.
A 2024 study published in the journal Nutrients suggests that AI provided “fairly good” nutritional guidance for some people with noncommunicable diseases, but its accuracy ranged between 55 and 73% in conditions such as fatty liver disease. However, when the robot was asked to provide advice to people suffering from more than one disease – such as diabetes and kidney disease together – contradictory and inappropriate advice appeared.
However, artificial intelligence is rapidly developing into an important source of general nutritional information. Another study published by the Journal of Nutrition and Metabolism in 2023 indicates that the performance of some models is equivalent to human performance when it comes to general advice, especially for those who cannot afford the cost of visiting a nutritionist to treat obesity, for example. In addition, obtaining an immediate answer at any time remains an advantage that artificial intelligence has over humans.
AI probably won’t replace the dietitian, but it will change the nature of their work: answering routine and common questions, while freeing up specialists’ time to handle complex cases, while simultaneously empowering users and increasing their participation in their health care.
Mistakes are never easy
According to a report by the New York Times, a young man in his thirties asked an artificial intelligence robot to help him develop a weight loss diet, so he suggested the keto diet, which is based on high fats and low carbohydrates. After weeks of following the diet, the young man remembered that his doctor had previously advised him to limit animal proteins because of his kidney stones. When he entered this information into the conversation, the bot asked him to immediately stop the keto diet, but that was after the potential harm had begun.
The flaw here is simple on the surface, but fundamental: the artificial intelligence did not ask him from the beginning about his medical history, it does not know what is not explicitly written for him, and it cannot conduct a clinical assessment as a doctor does.
In another study published by Frontiers in Nutrition, several artificial intelligence models were asked to create diets for a hypothetical group of obese teens. The robots prepared menus that contained, on average, 700 fewer calories than the dietitians for the same group. This significant difference means that if these teens follow these diets long-term, they may be at risk of eating disorders and malnutrition rather than healthy weight loss.
What does artificial intelligence get wrong nutritionally?
The most prominent problems with nutritional plans developed by artificial intelligence can be summarized in the following points:
- He does not recognize cases suffering from malnutrition well, and may provide advice that makes the problem worse.
- He does not screen for food allergies or eating disorder patterns, nor does he read body language or pick up danger signals like a doctor does in a clinic.
- High protein intake may be recommended for someone with kidney disease or liver problems, which poses real health risks.
- It does not place medical knowledge in a complex human and social context, but rather relies on linguistic patterns and guesses drawn from the data on which it is trained.
- Sometimes he tends to set unnecessarily strict regimens, or prices that are too low, especially for those who demand “quick” results.
- It sometimes produces information that appears accurate but is invented or undocumented, in what is known as an artificial intelligence “hallucinate.”
These errors do not mean that everything AI says is wrong, but they remind us that it is neither a personal doctor nor a certified nutritionist, but rather a tool that must be used with caution.
How do you use it to feed intelligently?
On busy lecture days or after a long day at work before going to the gym, an AI bot may be the closest option for getting a quick idea of the “perfect meal” or a healthy alternative to a favorite dish.
An article published by the Daily Orange website indicates that artificial intelligence tools in the field of nutrition are developing rapidly, but their capabilities are still uneven. They are able to explain the principles of a balanced diet, but they are not equipped to deal with complex diseases or sensitive conditions.
In general, experts advise using artificial intelligence in nutrition within clear limits, including:
- Request a simplified explanation of nutritional labels on foods.
- Get various ideas for healthy meals, or suggest lower-calorie alternatives to usual meals.
- Explore new food options, and understand the differences between alternatives such as types of fat or protein sources.
- Use it as a starting point for information, not an end point. That is, asking the robot about the sources, then verifying that they are reliable scientific sources.
- Do not rely on it to prescribe nutritional supplements or their dosages, or to develop treatment plans for complex medical conditions.
- Experience asking the same questions to more than one bot or tool; Different answers are an indication that there is something worth researching and investigating, and that relying on only one answer may be misleading.
Artificial intelligence remains, at best, a “digital nutritional assistant” that provides general information and preliminary ideas, not a substitute for a nutritionist who sees the patient, questions him, and understands his health history and daily life before writing him a diet that may affect his health for years.
