Patient says hospital used ChatGPT to write her treatment instructions
A woman with a rare genetic disorder says paperwork from her ER visit showed leftover ChatGPT search prompts, raising questions about undisclosed AI use in patient care.
What happened: A TikTok user named Tara Rule says she received an at-home treatment plan from a hospital that was apparently drafted with ChatGPT. In a video that went viral, Rule films the paperwork and points to what looks like a leftover browser search history printed at the top of the page, including prompts like 'Emergency Consultation Format' and 'Medical Spanish Interpreter.' Rule, who has Ehlers-Danlos Syndrome and other complex, interacting conditions, says the instructions she needed required more nuance than a generic chatbot prompt could provide. Neither the hospital nor the clinician involved has been publicly identified, and there is no independent confirmation beyond Rule's account and the document she shared, so details like which AI tool was used and why remain unclear.
Why it matters: Rule's case lands at a sensitive point in her medical history. She says she was once told she had brain cancer and only months to live, then spent years on treatments for conditions she didn't actually have, including medication for multiple sclerosis that she says triggered a dangerous reaction. A genetic test eventually diagnosed Ehlers-Danlos Syndrome, a connective-tissue disorder where standard treatment advice can be ineffective or even harmful. For someone with that history, generic aftercare instructions aren't just impersonal. They risk repeating the kind of one-size-fits-all mistakes that hurt her before, this time generated by software that never opened her chart.
What we don't know: This account comes from one TikTok video and the printed page shown in it, not from hospital records or an official statement. We don't know which AI tool was used, whether a clinician reviewed and approved the content before printing it, or whether the hospital has any policy on AI-assisted documentation. The visible search terms suggest someone used a chatbot to draft paperwork quickly, possibly for routine tasks like a translation request. It isn't confirmed that the chatbot wrote the actual medical guidance Rule describes receiving, rather than just generating boilerplate text around it.
The bigger pattern: Hospitals and clinics are increasingly experimenting with AI tools to draft notes, translate instructions, and speed up paperwork, often without clear rules on when patients must be told. Rule's video resonated because it captured a worry many patients already have: that stretched, under-resourced care systems might quietly hand complex judgment calls to a chatbot. Without hospital-level transparency policies or rules requiring disclosure, patients have little way to know whether a document in their hands came from a trained clinician, a chatbot, or some mix of both.