Crossmark

Multi-model assurance analysis showing large language models are highly vulnerable to adversarial hallucination attacks during clinical decision support

Crossref DOI link: https://doi.org/10.1038/s43856-025-01021-3

Published Online: 2025-08-02

Update policy: https://doi.org/10.1007/springer_crossmark_policy

Authors

Omar, Mahmud https://orcid.org/0009-0001-0438-0827
Sorin, Vera https://orcid.org/0000-0003-0509-4686
Collins, Jeremy D.

Reich, David

Freeman, Robert https://orcid.org/0000-0003-4946-6533
Gavin, Nicholas

Charney, Alexander https://orcid.org/0000-0001-8135-6858
Stump, Lisa

Bragazzi, Nicola Luigi https://orcid.org/0000-0001-8409-868X
Nadkarni, Girish N. https://orcid.org/0000-0001-6319-4314
Klang, Eyal https://orcid.org/0000-0002-4567-3108
License Information

Text and Data Mining valid from 2025-08-02

Version of Record valid from 2025-08-02
More Information

Article History

Received: 20 March 2025

Accepted: 10 July 2025

First Online: 2 August 2025

Competing interests

: The authors declare no competing interests.

: Was not required for this research as only synthetic open-access data was used.

Document is current

Any future updates will be listed below