# Taxonomy of Comprehensive Safety for Clinical Agents

arXiv: https://arxiv.org/abs/2509.22041 · HTML: https://mlsanity.com/paper/2509.22041/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 16 of its 41 references that are in the map.
- First posted on arXiv 26 Sep 2025.

## Builds on (most cited first, 16 of 16)

- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks](https://mlsanity.com/paper/2005.11401/) (2020)
- [Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedback](https://mlsanity.com/paper/2204.05862/) (2022)
- [Llama Guard: LLM-based Input-Output Safeguard for Human-AI Conversations](https://mlsanity.com/paper/2312.06674/) (2023)
- [Prompt Injection attack against LLM-integrated Applications](https://mlsanity.com/paper/2306.05499/) (2023)
- [NeMo Guardrails: A Toolkit for Controllable and Safe LLM Applications with Programmable Rails](https://mlsanity.com/paper/2310.10501/) (2023)
- [Introducing v0.5 of the AI Safety Benchmark from MLCommons](https://mlsanity.com/paper/2404.12241/) (2024)
- [Conformal Prediction for Natural Language Processing: A Survey](https://mlsanity.com/paper/2405.01976/) (2024)
- [KR-BERT: A Small-Scale Korean-Specific Language Model](https://mlsanity.com/paper/2008.03979/) (2020)
- [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)
- [Intent Identification and Entity Extraction for Healthcare Queries in Indic Languages](https://mlsanity.com/paper/2302.09685/) (2023)
- [Enhancing Guardrails for Safe and Secure Healthcare AI](https://mlsanity.com/paper/2409.17190/) (2024)
- [Self-Diagnosis through AI-enabled Chatbot-based Symptom Checkers: User Experiences and Design Considerations](https://mlsanity.com/paper/2101.04796/) (2021)
- [Transfer Learning Robustness in Multi-Class Categorization by Fine-Tuning Pre-Trained Contextualized Language Models](https://mlsanity.com/paper/1909.03564/) (2019)
- [The Right Model for the Job: An Evaluation of Legal Multi-Label Classification Baselines](https://mlsanity.com/paper/2401.11852/) (2024)
- [Specialty detection in the context of telemedicine in a highly imbalanced multi-class distribution](https://mlsanity.com/paper/2402.14039/) (2024)

## Abstract (arXiv)

Safety is a paramount concern in clinical chatbot applications, where inaccurate or harmful responses can lead to serious consequences. Existing methods--such as guardrails and tool calling--often fall short in addressing the nuanced demands of the clinical domain. In this paper, we introduce TACOS (TAxonomy of COmprehensive Safety for Clinical Agents), a fine-grained, 21-class taxonomy that integrates safety filtering and tool selection into a single user intent classification step. TACOS is a taxonomy that can cover a wide spectrum of clinical and non-clinical queries, explicitly modeling varying safety thresholds and external tool dependencies. To validate our taxonomy, we curate a TACOS-annotated dataset and perform extensive experiments. Our results demonstrate the value of a new taxonomy specialized for clinical agent settings, and reveal useful insights about train data distribution and pretrained knowledge of base models.