Story perspectives

Oxford Study: AI Misdiagnoses, Can't Replace Doctors Yet

2/19/2026

Summary

Story summary

An Oxford University study warns against using large-language models (LLMs) for medical advice after testing 1,300 participants seeking guidance from ChatGPT.
The results show LLMs often give inaccurate medical advice due to communication breakdowns.
Dr. Rebecca Payne said AI is not ready to replace physicians, as it can misdiagnose or miss urgent needs.
A separate study finds LLMs can be easily programmed to provide false information, fueling disinformation concerns.

1 of 1

Story summary

An Oxford University study warns against using large-language models (LLMs) for medical advice after testing 1,300 participants seeking guidance from ChatGPT.
The results show LLMs often give inaccurate medical advice due to communication breakdowns.
Dr. Rebecca Payne said AI is not ready to replace physicians, as it can misdiagnose or miss urgent needs.
A separate study finds LLMs can be easily programmed to provide false information, fueling disinformation concerns.