A 60-year-old man from New York was hospitalized after following a strict salt-reduction plan suggested by ChatGPT, resulting in hyponatraemia due to dangerously low sodium levels. He cut sodium to nearly zero without consulting a doctor, relying solely on AI-generated advice. The case, published in the American College of Physicians journal, underscores the dangers of heeding AI health recommendations without professional oversight. ChatGPT suggested sodium bromide as a salt substitute, which led to bromide toxicity in the patient. He suffered from hallucinations, confusion, and other neurological symptoms, necessitating a three-week hospital stay for rehydration and electrolyte balance restoration. This incident amplifies concerns about health misinformation from AI tools, which can inadvertently spread inaccuracies. OpenAI specifies in its Terms of Use that the AI should not serve as a sole source of medical advice. As AI use increases, the importance of critical thinking and expert consultation becomes paramount for health-related matters.
Source link
