Man hospitalised with hallucinations after reportedly following ChatGPT diet advice to replace salt with sodium bromide
A 60-year-old man looking to cut chloride from his diet reportedly followed dietary advice from ChatGPT and replaced table salt with a chemical called sodium bromide. He later developed hallucinations and paranoia and needed hospital treatment. The chatbot conversation could not be retrieved, so doctors could not determine exactly what advice he had received. The case is described in a paper by researchers from Binghamton, Stanford, Texas A&M and Indiana universities that calls for greater transparency over AI health advice. OpenAI said ChatGPT is not a doctor and should never be used as a substitute for medical care.
What the AI did
ChatGPT reportedly gave dietary advice to a man who wanted to cut chloride from his diet, and he went on to replace table salt with sodium bromide. Exactly what ChatGPT told him is not known, because the conversation could not be retrieved.
First reported October 18, 2026 · Added to the register October 30, 2026 · 1 source
- Developer
- OpenAI
- Model
- ChatGPT
- Affected
- A 60-year-old man
- When it happened
- Not stated in the sources
- First reported
- October 18, 2026
What this means for you
Could this affect you?
If you offer chatbots for health-related use, or your people rely on them for health decisions. Consumers who use general-purpose chatbots for diet, medication or treatment decisions are exposed, and so are health-care providers who may be unable to reconstruct the AI advice a patient followed.
What to check
- Keep retrievable records of chatbot conversations used for health-related questions.
- Add clear medical disclaimers and prompts directing users to a clinician.
- Provide a way for users to report harmful chatbot answers.
- Verify AI health guidance with a clinician before changing diet or medication.
Areas of your AI programme this touches
Every fact and its source (5)
- Evidence gapChatbot conversation could not be retrieved“the chatbot conversation could not be retrieved, doctors were unable to determine exactly what advice the man had received”[1]
- Paper authors' institutionsBinghamton, Stanford, Texas A&M and Indiana universities“researchers from Binghamton University, Stanford University, Texas A&M University and Indiana University”[1]
- OpenAI responseChatGPT not a substitute for medical care“ChatGPT is not a doctor and should never be used as a substitute for medical care, diagnosis, or treatment.”[1]
- Substance used in place of saltSodium bromide“replaced table salt with a chemical compound called sodium bromide”[1]
- Health outcomeHallucinations, paranoia and hospital treatment“He later developed hallucinations and paranoia and required hospital treatment”[1]
Sources
- A Man Turned to AI for Diet Advice. His Hospital Stay Exposed a Blind Spotnewsweek.com · October 18, 2026
How this record is classified. Severity M (3/5): harm to individuals, or a policy breach with limited reach. OECD level: incident, an event in which an AI system led to actual harm. Evidence: Reported, meaning press reporting only.
The facts in At a glance and in Every fact and its source are quoted from the sources listed. The summary, What the AI did and What this means for you are written from those sources and checked against them automatically before publication. Records reflect what has been disclosed, not everything that has happened. Incidents are found through news feeds and the GDELT Project.
