Ask a medical question to GBT Chat, Gemini, or any other AI-powered chatbot, and you may be amazed at how quickly it can provide a detailed answer. He can formulate the answer to suit your needs according to the commands you direct to him. However, you have to be careful if you want to truly benefit from this technology.
While AI in healthcare has impressive capabilities, it also has limitations, especially when applied by patients themselves, rather than doctors. In this regard, Dr. Adam Rodman, a general physician in internal medicine and an artificial intelligence researcher at Beth Israel Deaconess Medical Center, affiliated with Harvard University, said: “AI-powered chatbots are good for some medical inquiries but not others.”
To help people understand how best to exploit AI in healthcare, Dr. Rudman developed a traffic light system; A green light appears for safe and useful uses, while questions with a yellow light require caution when asked, while a red light identifies inquiries that are best avoided completely.
Artificial Intelligence is rapidly becoming an indispensable tool in all medical fields, as it shapes the way doctors and other health professionals work, from basic research to diagnosing and treating patients.
AI systems learn patterns from huge amounts of data, and then use these patterns to perform tasks typically associated with human thinking. Large language models are a type of artificial intelligence that can recognize and generate human language in a manner similar to human language. AI systems can also undertake tasks, such as identifying signs of cancer in mammograms, that would be difficult for humans to detect.
A growing number of doctors are using AI tools, including large language models, to help make clinical decisions, take notes in medical records, and keep up with medical knowledge.
“The 'green light' questions are the ones you can comfortably ask without discussing them with your doctor, that is, your health care provider,” Dr. Rodman said. For example, you can ask a chatbot to help you understand general medical information, such as: “Explain diabetes to me,” or “What is high blood pressure?” You may also ask the chatbot for advice about diet, such as: “I suffer from lactose intolerance, can you help me develop a suitable menu plan?”
Chatbots can also explain the meaning of doctors' notes, lab reports, and imaging test results. If you don't understand what's written in your medical record, you can ask the chatbot to "explain what the doctor meant by this note" (removing personal identifiers first).
In addition, the AI-powered chatbot can also help you prepare for a doctor's visit; You can enter information from your last visit and ask: “How should I prepare for this visit?” Tell me three ways.
You can ask “yellow light” questions, but keep in mind that the answers will not replace a doctor's clinical assessment. People often ask chatbots about new symptoms they are experiencing. “There are studies that suggest that large AI-powered language models are just as good, if not better than humans, at detecting new symptoms,” Dr. Rudman explained. However, the fact remains that the level of accuracy depends on the context.
Answers can reflect the way you interact with the chatbot. “For example, these robots will often tell you what they discover you want to hear,” Dr. Rodman said. “Which could be dangerous.” This may lead to misinformation and anxiety.
And so; Instead of writing out a list of symptoms, Dr. Rodman recommended using a command, such as: “I'm having a new headache.” "Can you interview me, as if you were a doctor, to help me explore my symptoms?" Here, the chatbot will proceed to ask questions that put your symptoms in context. In addition to the symptoms you are experiencing, he will ask about the symptoms you are not experiencing.
Dr. Rodman added: “While you can explore new symptoms and potential diagnoses using a chatbot, this should only be done as a first step before talking to a healthcare professional.”
Based on the current limitations of AI; You cannot rely on the accuracy of answers on how to treat and manage health conditions. Avoid asking questions like: Is this treatment plan right for me? Should I take this medication? Is this chemotherapy regimen the right one? These are questions that could be classified as “red light” questions.
Regarding this, Dr. Rodman said: “I tell patients, simply don’t do it.” Treatment decisions involve so many considerations, including physical examinations, medical history, and clinical judgment, that trusting a chatbot is not foolproof.
By the time you read this topic, AI chatbots will have evolved – that's simply how quickly technology changes. However, this does not change the basic takeaway: you should not rely on a medical diagnosis provided by artificial intelligence, and do not consider a chatbot your personal AI doctor. “AI is not a substitute for healthcare professionals,” Dr. Rodman explained.
Doctors are increasingly concerned that weight-loss medications are contributing to new cases of anorexia nervosa, and that there is a lack of safeguards to protect patients with eating disorders.
Tens of millions of patients are currently taking these medications, which can sometimes cause symptoms similar to those of anorexia nervosa: hair loss, dizziness, digestive problems, and severe calorie deficiency. “In this case, we wonder: Is this anorexia nervosa?” says Dr. Katherine Kelly, a family and sports medicine physician in Long Beach, California. Is it an eating disorder? Or is it just a side effect of the medication?
Citing an “urgent need” for guidance, an international coalition of obesity medicine specialists, eating disorder doctors, and other experts recently issued new recommendations for prescribing glucagon-like peptide-1 (GLP-1) agonists and eating disorders. Among their suggestions for doctors: Screen all patients for signs of eating disorders before starting to take glucagon-like peptide receptor agonists, and monitor anyone taking these medications periodically.
Neither of these procedures is currently approved practice. Also, most doctors are not trained to recognize signs of eating disorders, which may limit the scope of the new recommendations.
What makes matters more complicated is that many people are obtaining these medications via telemedicine services, with limited or no medical supervision. In a recent study, researchers who tried to purchase weight-loss medications from 49 websites found that only a third of these sites required direct communication with a doctor before prescribing the medications, and that only half of them asked about eating disorders.
“We're starting to prescribe a very dangerous drug when we don't fully know all the side effects,” said Vanessa Longley, chief executive of Beat, a British non-profit that runs a hotline to help people with eating disorders.
There is not enough data available on how common anorexia nervosa is among people who take weight-loss medications. But a recent study examining the use of GLP-1 among people without any apparent medical condition found that the rate of eating disorders in this group was six times higher than the rate in the general population. Research indicates that many people are turning to these medications to achieve dangerous levels of weight loss. If left untreated, anorexia nervosa can lead to osteoporosis, bone fractures, cardiovascular problems, and cognitive disorders. It may also be fatal.
Even if a patient discloses his or her history of anorexia nervosa, doctors do not have adequate guidance about the next step. Do they prescribe medication anyway? Or do they refrain from prescribing it because of this risk, abandoning its benefits?
A growing body of research suggests that for some people with a genetic predisposition to anorexia nervosa, a calorie deficit — such as going on a diet for the first time — may be the trigger for an eating disorder, as their brains are primed to view this calorie deficit as particularly satisfying, creating a vicious cycle that prompts them to further restrict their food intake. The researchers also note that obesity surgery may cause loss of appetite in this way.
There is reason to believe that something similar may happen to some patients taking GLP-1 drugs. “This is a critical turning point,” said Cynthia Bulik, founding director of the Center of Excellence for Eating Disorders at the University of North Carolina and a member of the expert panel that issued the new guidelines. “It stimulates their latent genetic predisposition.”
Dr. Amanda Banks, a physician in Philadelphia and author of a book about the relationship between anorexia nervosa and weight-loss drugs, described these drugs as “powerful pharmacological tools” for suppressing appetite. This makes it an effective treatment for people suffering from obesity. It is also one of the reasons why these medications show promising results in patients with binge eating disorder.
Doctors sometimes diagnose anorexia nervosa by its physical symptoms: rapid and severe weight loss, dehydration, dizziness, and hair loss caused by poor nutrition. But these symptoms may also be the effects of the hormones of GLP-1 medications, which makes diagnosing an eating disorder difficult.
It is also difficult for doctors to distinguish between whether the patient is deliberately trying to reduce the amount of food he eats, or whether the medication is what is pushing him to do so. There is no definitive answer as to how much weight loss is excessive when taking medications designed to help people lose weight.
“We're seeing patients with what we call pathological weight loss,” which can mean losing an unhealthy amount of weight, said Dr. Aaron Kishen, a Nova Scotia-based eating disorder specialist who co-authored the new guidelines. He added: “And they take GLP-1 hormones.”
The distinction between the use of GLP-1 hormones and eating disorders has become so blurred that some researchers question whether the clinical definition of anorexia nervosa needs to change to reflect the symptoms of this disorder in those taking these medications.
While the new guidelines are an attempt to protect against the risks of GLP-1 drugs that may harm patients with eating disorders, implementing these recommendations in practice may be a challenge, especially with the increase in online prescriptions. "If any of us can say, 'I'm 165 pounds, and I have diabetes and high blood pressure,' and none of those things are true, and no one checks that, and then they can get the medication at their door in the morning, that's a problem," Banks says.
Even doctors who provide direct care to patients taking these medications are not in a position to address the risks of eating disorders. Primary care doctors, who frequently prescribe these medications, are often strapped for time and must treat a variety of health problems.
Obesity doctors and eating disorder specialists who can help address these issues with patients rarely have opportunities to interact with each other, says Dr. Melanie Jay, director of NYU Langone's Comprehensive Obesity Program. Obesity doctors lack adequate guidance on how to treat anorexia nervosa, or even how to spot its signs in the first place.
“The biggest risk is not the medication, it's not diagnosing your underlying eating disorder,” said Longley, president of the British Eating Disorders Non-Profit.
A single reading may reveal a fleeting change, but what do hundreds of readings tell us? Between the promise of prevention and the worry of every spike, the proliferation of wearable devices raises a question beyond sensor accuracy: Does monitoring our bodies moment-by-moment give us a greater ability to avoid illness? Or does it make us chase numbers that we don't always know how to interpret?
This question emerged in the session “Wearable Devices, Glucose Monitoring and the Prevention Economy” within the Digital Forum of the Global Health Forum. After rescheduling the in-person exhibition to October 25-28, 2027, the organizers launched a digital forum that extends from October 26, 2026 to December 20, 2026.
The Digital Forum is not limited to moving exhibition sessions online; Its program includes 30 professional seminars and 15 dialogues with health leaders, in addition to a report on the health care market. Its importance lies in maintaining discussion among specialists, investors and decision-makers, and providing access to topics of digital health and medical innovation beyond the boundaries of the venue and the days of the exhibition. It also enhances the presence of the Kingdom of Saudi Arabia on the global health scene, and highlights its ambition to consolidate its leadership in employing artificial intelligence and digital technologies in health care. But it remains complementary to face-to-face meetings, with the face-to-face demonstration of techniques and building partnerships they provide.
These discussions come in the context of the expansion of the Global Health Forum itself. In the 2025 edition, it attracted more than 130,000 visitors, more than 2,200 brands were displayed, and 607 speakers participated, while the organizers announced agreements and deals worth more than $35.5 billion. These numbers reflect the expansion of the forum from an exhibition of devices and services to a platform where questions of investment, innovation, and the future of health care intersect.
Continuous glucose monitors have become an important tool for many people with diabetes. A small sensor is placed under the skin; It measures glucose in interstitial fluid, then sends its readings to a phone or receiver. Unlike a finger prick, which gives a reading in a single moment, these devices monitor the change in glucose levels over time, helping the patient monitor their trend, notice whether they are rising or falling, and discuss the impact of food, activity, and treatment with the doctor.
But the use of the technology has expanded beyond monitoring people with diabetes, and some healthy people have turned to it in an effort to “improve performance” or achieve ideal health. Here the question differs: What do these readings mean for those who do not suffer from diabetes? Is monitoring glucose changes enough to change habits that actually improve health?
On January 26, 2026, the European Journal of Medical Research published a scientific review that covered 23 studies that included 1,074 people without diabetes, and statistically collected the results of the studies that were able to compare them. The analysis showed a lower average glucose level in device users compared to participants who did not use them, while their use was associated with dietary changes and greater adherence to health interventions.
But the results varied depending on the participants; There was no significant difference in body mass index, and no clear improvement in glucose control was observed in healthy people with normal levels, while the results seemed more promising in those with prediabetes. The review indicates that the devices may help within a structured lifestyle change program, rather than as a standalone treatment. The short duration of many studies, and some of them combining the device with health education, make it difficult to determine the effect of the sensor alone.
A randomized trial published by the New England Journal of Medicine in 2026 illustrates the difference between using technology to monitor an existing disease and generalizing it to healthy people. The trial included 283 adults with type 2 diabetes who do not use insulin. After 26 weeks, users of continuous monitoring devices had lower blood sugar by an additional 0.9 percentage points compared to those who received usual care, and they also spent more time within the target range.
Although this is an important finding for these patients; But it does not prove that the device prevents diabetes in healthy people. The category and research question are different. The device measures glucose; But it does not diagnose the disease alone. A temporary rise after a meal is not enough to diagnose a disorder, and a normal fluctuation does not mean that the body has lost its ability to regulate sugar.
