Medical large language models are vulnerable to data-poisoning attacks

Daniel Alexander Alber; Zihao Yang; Anton Alyakin; Eunice Yang; N. Shesh; Aly Valliani; Jeff Zhang; Gabriel R. Rosenbaum; Ashley K. Amend-Thomas; David B. Kurland; C. Kremer; Alexander Eremiev; Bruck Negash; Daniel D. Wiggan; M. Nakatsuka; Karl L. Sangwon; Sean N. Neifert; Hammad A. Khan; Akshay Save; Adhith Palla; Eric A. Grin; Monika Hedman; Mustafa Nasir-Moin; Xujin Chris Liu; Lavender Yao Jiang; Michal Mankowski; Dorry L. Segev; Yindalon Aphinyanaphongs; Howard A. Riina; John G. Golfinos; Daniel A. Orringer; Douglas Kondziolka; Eric K. Oermann

doi:10.1038/s41591-024-03445-1

JOURNAL ARTICLE

Medical large language models are vulnerable to data-poisoning attacks

Daniel Alexander Alber Zihao Yang Anton Alyakin Eunice Yang N. Shesh Aly Valliani Jeff Zhang Gabriel R. Rosenbaum Ashley K. Amend-Thomas David B. Kurland C. Kremer Alexander Eremiev Bruck Negash Daniel D. Wiggan M. Nakatsuka Karl L. Sangwon Sean N. Neifert Hammad A. Khan Akshay Save Adhith Palla Eric A. Grin Monika Hedman Mustafa Nasir-Moin Xujin Chris Liu Lavender Yao Jiang Michal Mankowski Dorry L. Segev Yindalon Aphinyanaphongs Howard A. Riina John G. Golfinos Daniel A. Orringer Douglas Kondziolka Eric K. Oermann

Year: 2025 Journal: Nature Medicine Vol: 31 (2)Pages: 618-626 Publisher: Nature Portfolio

DOI: 10.1038/s41591-024-03445-1

Get Full-Text PDF Get Analytical Report

Keywords:

Misinformation Computer science Harm The Internet Internet privacy Computer security Health care Data science Psychology World Wide Web Political science

Metrics

Cited By

194.21

FWCI (Field Weighted Citation Impact)

Refs

1.00

Citation Normalized Percentile

Is in top 1%

Is in top 10%

Citation History

Topics

Artificial Intelligence in Healthcare and Education

Health Sciences → Medicine → Health Informatics

Topic Modeling

Physical Sciences → Computer Science → Artificial Intelligence

COVID-19 diagnosis using AI

Health Sciences → Medicine → Radiology, Nuclear Medicine and Imaging

Medical large language models are vulnerable to data-poisoning attacks

Metrics

Citation History

Topics

Related Documents

Training Time Vulnerabilities in Large Language Models: Data Poisoning and Backdoor Attacks

Poisoning medical knowledge using large language models

Understanding and Mitigating Poisoning Attacks in Large Language Models

Understanding and Mitigating Poisoning Attacks in Large Language Models

Medical MLLM Is Vulnerable: Cross-Modality Jailbreak and Mismatched Attacks on Medical Multimodal Large Language Models