A 60?year?old man found himself hospitalised after three months of following ChatGPT’s recommendation to replace table salt with sodium bromide (bromism risk)
In turn, this causes paranoia and hallucinations. The initial advice proved dangerously misleading.
As more users in Malaysia and beyond turn to AI for health tips, this case underscores why accuracy is vital — especially when one “quick fix” can cause real harm. Our take? Let AI guide, but double?check with a pro.
What we could expect
The Annals of Internal Medicine: Clinical Cases published this unusual case on 5 August 2025. It involved a man seeking to eliminate chloride after reading about the harms of sodium chloride. Consulting ChatGPT (likely versions 3.5 or 4.0) led him to believe sodium bromide was an acceptable substitute
For three months, he consumed sodium bromide he bought online. Though bromide was once used in early sedatives and OTC sleep remedies of the 19th and 20th centuries, it has long been banned due to its toxicity
Gizmodo
Symptoms set in gradually, starting with confusion and paranoia, evolving into hallucinations. Lab tests flagged elevated chloride, but doctors later identified “pseudohyperchloremia” caused by bromide interference
His condition escalated to psychosis and required an involuntary psychiatric hold. Treatment involved IV fluids, electrolytes, and antipsychotics. Over three weeks, he recovered and was stable at a two?week follow?up
OpenAI reaffirmed that ChatGPT is not intended for medical diagnosis or treatment and urged users to consult a professional, a caution echoed by the case study authors
Think twice before relying solely on AI for health decisions. Have you or someone you know tried following AI?based advice? Share your experience below.
Stay tuned to TechNave.com for more updates.
COMMENTS