A Florida pastor has filed a lawsuit against OpenAI, alleging that ChatGPT dissuaded him from seeking medical treatment for a pulmonary embolism, a condition with a mortality rate of 30%. The case, reported by the New York Times, highlights the dangers of relying on artificial intelligence for health advice.
ChatGPT's dangerous advice worsened the patient's condition
According to the complaint, Scott Winters, a 47-year-old pastor, consulted ChatGPT about concerning symptoms. The chatbot, based on the GPT-4o model, allegedly assured him that the symptoms were not dangerous and advised him to rest, trusting that God did not design the human body to fail. Instead of seeing a doctor, Winters followed the advice, remaining immobile for weeks. When he was finally admitted to intensive care, doctors diagnosed a series of pulmonary embolisms exacerbated by prolonged inactivity.
This incident is not isolated. Hundreds of millions of people ask AI chatbots health questions every week. An analysis of searches on Microsoft Copilot found health queries among the most common. Despite warnings about AI hallucinations, many users continue to treat these tools as substitutes for doctors.
Sponsored Protocol
The risks of relying on AI for medical diagnoses
The American Medical Association and other health organizations have repeatedly cautioned against using chatbots for medical advice. AI models lack the ability to perform physical exams or assess symptom complexity. Moreover, they tend to generate confident yet incorrect responses, a phenomenon known as hallucination. In this case, the chatbot even introduced a religious reference that reinforced the pastor's decision to avoid medical care.
OpenAI has acknowledged the lawsuit and stated it is working to improve model safety, but the damage is done. This episode demonstrates that, however useful in many areas, chatbots should never replace professional medical advice. As experts note, technology can support diagnosis but not substitute human clinical judgment.
Sponsored Protocol
For reliable health information, sources like the World Health Organization or the Centers for Disease Control are preferable. Even a simple Google search can be safer, as search engines at least cite sources. In an era where algorithms increasingly match data, as discussed in the article about native JavaScript Fetch API for handling data, it is vital to remember that medical data requires careful human oversight.
The case of Pastor Winters serves as a warning. As AI evolves, caution must remain the rule when health is at stake. No algorithm can replace a doctor's advice.
