The risks of diagnosis without physical examination While AI tools like ChatGPT have become popular for information retrieval, medical professionals are raising alarms about patients using these platforms for self-diagnosis and treatment. A fundamental flaw of AI in medicine is its reliance on subjective user descriptions (subjective symptoms) without the ability to perform physical examinations (objective signs). The lack of direct clinical observation, diagnostic imaging, and laboratory tests often leads to inaccurate or dangerously delayed diagnoses.
Documented cases show that following AI-generated advice has led to severe consequences, including patients missing the “golden hour” for stroke intervention or suffering from chemical poisoning due to incorrect dietary recommendations provided by chatbots.
Information vs. Clinical indication: A critical distinction Medical experts emphasize that healthcare must be “individualized.” A medication that is safe for one individual could be life-threatening for another due to underlying conditions, allergies, or complex drug interactions. Current AI models lack the capacity to comprehensively analyze a patient’s medical history, genetic profile, or organ function to determine a safe and effective dosage.
Furthermore, purchasing prescription drugs—especially antibiotics—based on AI suggestions contributes significantly to the global crisis of antibiotic resistance and risks masking serious conditions like pneumonia or acute infections.
Professional recommendations
-
AI is not a doctor: Chatbots should only be used for general information, never as a substitute for professional medical advice, diagnosis, or treatment.
-
Prioritize emergency care: In life-threatening situations such as chest pain, difficulty breathing, or sudden loss of consciousness, individuals should contact emergency services immediately rather than consulting AI.
-
Consult specialists: Any concerns regarding medication side effects must be discussed with qualified healthcare providers to ensure adjustments are made based on clinical evidence and professional ethics.
Source: https://tuoitre.vn/nguy-hiem-tu-don-thuoc-chatgpt-20251230080644576.htm