Agreement among large language models does not track medical truth in health claim verification

A study published in Nature reveals that consensus among large language models does not correlate with medical accuracy. Researchers found that even when models agree, they frequently provide incorrect or disputed health information, a phenomenon they call the 'consensus mirage'.
Why it matters
This highlights significant risks in relying on AI for medical advice, suggesting that model agreement is not a reliable proxy for truth in high-stakes domains.
Scientific Reports ( 2026 ) Cite this article
We’re sharing this article early to provide faster access to peer-reviewed, accepted research. It is citable and carries a permanent DOI. This version is subject to further edits and will be replaced automatically by the final Version of Record. All legal disclaimers apply.
Get smarter about the news
Sign up free for a feed built around what you actually care about, Dive Deeper research on any story, and the full text of every article.
Create free accountAlready have an account? Sign in