According to AIbase, a exclusive report provided by OpenAI to Axios reveals an astonishing trend: healthcare has become one of the core use cases for ChatGPT. More than 5% of queries worldwide are related to health, and in the United States, about 40 million people seek medical advice from this AI every day. From deciphering complicated medical bills, comparing insurance plans, to preliminary symptom checking, ChatGPT is becoming a "medical ally" for many Americans who cannot see a doctor immediately.

OpenAI

OpenAI clearly recognized this market demand and has positioned its next-generation model, GPT-5, as a specialized intelligent agent for the healthcare sector. Compared to previous versions, GPT-5 shows stronger expertise in handling complex clinical reasoning and interpreting medical images such as pathology reports and MRIs. The report also notes that ChatGPT currently handles nearly 2 million insurance-related questions each week. This surge in demand is driven by the fact that starting from the beginning of 2026, the Trump administration allowed long-standing Medicare subsidies to expire, leading to significant premium increases for millions of Americans, forcing them to turn to AI for more cost-effective alternatives.

However, relying on AI for medical advice still carries significant risks. Although GPT-5 has made progress in reducing hallucinations, the characteristic of the model to fabricate facts has not been completely eliminated. Experts warn that the risk is higher when users consult in voice mode, as voice mode often uses lighter models with weaker reasoning capabilities to ensure response speed, which could lead to misdiagnosis or missed diagnoses. Despite OpenAI's latest promotion highlighting its potential as a "medical assistant," the company has not adequately warned users about the possible hallucinations of the model and its performance limitations in low-power modes.