Medical large language models are vulnerable to data-poisoning attacks

By D. Alber, Zihao Yang, A. Alyakin, E. Yang, Sumedha Rai, Aly A. Valliani, Jeff Zhang, Gabriel R. Rosenbaum, Ashley K. Amend-Thomas, David B. Kurland and 23 more. 2025

Subfields: Large Language Models, Safety, Robustness & Interpretability, Medical & Health AI

Citations in the corpus

• General-purpose large language models outperform specialized clinical AI tools on medical benchmarks (2026, unknown): “The path forward may ultimately lie with hospital-specific LLMs that leverage institutional data 13,14 to mitigate external harm 15 , along with careful use of frontier models for less-sensitive tasks 16 .”
• Evaluating the Performance and Fragility of Large Language Models on the Self-Assessment for Neurological Surgeons (2025, unknown): “Recent work has shown that LLMs can fail under subtle perturbations, 17 variation in clinical note type, 24 training data-poisoning, 25 or even internal tampering, 26 and that fi ne-tuning alone does not confer resistance to distractions.”