Man develops rare condition after ChatGPT query over stopping eating salt

Summary

A US medical journal has warned against using ChatGPT for health information after a man developed a rare condition following an interaction with the chatbot about removing table salt from his diet.An article in the Annals of Internal Medicine reported a case in which a 60-year-old man developed bromism, also known as bromide toxicity, after consulting ChatGPT.The article described bromism as a “well-recognised” syndrome in the early 20th century that was thought to have contributed to almost one in 10 psychiatric admissions at the time.The patient told doctors that after reading about the negative effects of sodium chloride, or table salt, he consulted ChatGPT about eliminating chloride from his diet and started taking sodium bromide over a three-month period. This was despite reading that “chloride can be swapped with bromide, though likely for other purposes, such as cleaning”. Sodium bromide was used as a sedative in the early 20th century.The article’s authors, from the University of Washington in Seattle, said the case highlighted “how the use of artificial intelligence can potentially contribute to the development of preventable adverse health outcomes”.They added that because they could not access the patient’s ChatGPT conversation log it was not possible to determine the advice the man had received.Nonetheless, when the authors consulted ChatGPT themselves about what chloride could be replaced with, the response also included bromide, did not provide a specific health warning and did not ask why the authors were seeking such information – “as we presume a medical professional would do”, they wrote.The authors warned that ChatGPT and other AI apps could ‘“generate scientific inaccuracies, lack the ability to critically discuss results, and ultimately fuel the spread of misinformation”.The company announced an upgrade of the chatbot last week and claimed one of its biggest strengths was in health. It said ChatGPT – now powered by the GPT-5 model – would be be...