# When Medical Safety Alignment Fails: A Benchmark for Evaluating LLMs on High-Risk Medical Queries

arXiv: https://arxiv.org/abs/2606.28332 · HTML: https://mlsanity.com/paper/2606.28332/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Safety alignment, refusal and guardrails, one of 12 directions in LLM safety and security.
- Builds on 11 of its 21 references that are in the map.
- First posted on arXiv 26 May 2026.

## Placed in

- LLM safety and security / [Safety alignment, refusal and guardrails](https://mlsanity.com/safety/safety-alignment-refusal-and-guardrails/)

## Builds on (most cited first, 11 of 11)

- [Training language models to follow instructions with human feedback](https://mlsanity.com/paper/2203.02155/) (2022)
- [What Disease does this Patient Have? A Large-scale Open Domain Question Answering Dataset from Medical Exams](https://mlsanity.com/paper/2009.13081/) (2020)
- [Llama Guard: LLM-based Input-Output Safeguard for Human-AI Conversations](https://mlsanity.com/paper/2312.06674/) (2023)
- [Fine-tuning Aligned Language Models Compromises Safety, Even When Users Do Not Intend To!](https://mlsanity.com/paper/2310.03693/) (2023)
- [A Novel Evaluation Benchmark for Medical LLMs: Illuminating Safety and Effectiveness in Clinical Domains](https://mlsanity.com/paper/2507.23486/) (2025)
- [CARES: Comprehensive Evaluation of Safety and Adversarial Robustness in Medical LLMs](https://mlsanity.com/paper/2505.11413/) (2025)
- [Rethinking Safety in LLM Fine-tuning: An Optimization Perspective](https://mlsanity.com/paper/2508.12531/) (2025)
- [LiveMedBench: A Contamination-Free Medical Benchmark for LLMs with Automated Rubric Evaluation](https://mlsanity.com/paper/2602.10367/) (2026)
- [The Need for Guardrails with Large Language Models in Medical Safety-Critical Settings: An Artificial Intelligence Application in the Pharmacovigilance Ecosystem](https://mlsanity.com/paper/2407.18322/) (2024)
- [Enhancing Guardrails for Safe and Secure Healthcare AI](https://mlsanity.com/paper/2409.17190/) (2024)
- [Improving the Safety and Trustworthiness of Medical AI via Multi-Agent Evaluation Loops](https://mlsanity.com/paper/2601.13268/) (2026)

## Abstract (arXiv)

Large language models (LLMs) are increasingly used for medical and health-related questions, yet their safety in high-risk medical scenarios remains poorly understood. We introduce textscMedHarmfootnoteCode and data will be released upon acceptance. Due to the sensitive nature of high-risk medical queries, data access will be available to qualified researchers upon request., a high-risk medical safety benchmark with 1,100 medically grounded queries across 10 safety-critical categories, including toxicology, pharmacology, covert poisoning, anesthesia, and fetal harm. Unlike broad medical QA benchmarks, textscMedHarm targets realistic clinical, educational, and technical prompts that require refusal, caution, or safe redirection rather than direct helpfulness. We evaluate 15 LLMs spanning general-purpose, medical-purpose, closed-source, and downstream SFT models, together with 4 representative guardrail models. Results reveal a substantial gap between apparent alignment and medical safety: aligned models can still produce unsafe or actionable responses, medical fine-tuning can amplify harmful specificity, and external guardrails reduce some failures while introducing brittle blocking and weak safe helpfulness. These findings show that medical safety cannot be inferred from general alignment or medical capability alone, highlighting the need for domain-specific stress testing before deploying LLMs in safety-critical medical applications.