# LabSafety Bench: Benchmarking LLMs on Safety Issues in Scientific Labs

arXiv: https://arxiv.org/abs/2410.14182 · HTML: https://mlsanity.com/paper/2410.14182/

## Key facts

- Cited by 25 papers in the map (data through 14 Oct 2026).
- Builds on 13 of its 37 references that are in the map.
- First posted on arXiv 18 Oct 2024.

## Builds on (most cited first, 13 of 13)

- [GPT-4 Technical Report](https://mlsanity.com/paper/2303.08774/) (2023)
- [Llama 2: Open Foundation and Fine-Tuned Chat Models](https://mlsanity.com/paper/2307.09288/) (2023)
- [Chain-of-Thought Prompting Elicits Reasoning in Large Language Models](https://mlsanity.com/paper/2201.11903/) (2022)
- [Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks](https://mlsanity.com/paper/2005.11401/) (2020)
- [Judging LLM-as-a-Judge with MT-Bench and Chatbot Arena](https://mlsanity.com/paper/2306.05685/) (2023)
- [Pre-train, Prompt, and Predict: A Systematic Survey of Prompting Methods in Natural Language Processing](https://mlsanity.com/paper/2107.13586/) (2021)
- [A Survey on Hallucination in Large Language Models: Principles, Taxonomy, Challenges, and Open Questions](https://mlsanity.com/paper/2311.05232/) (2023)
- [ChemCrow: Augmenting large-language models with chemistry tools](https://mlsanity.com/paper/2304.05376/) (2023)
- [What can Large Language Models do in chemistry? A comprehensive benchmark on eight tasks](https://mlsanity.com/paper/2305.18365/) (2023)
- [SciEval: A Multi-Level Large Language Model Evaluation Benchmark for Scientific Research](https://mlsanity.com/paper/2308.13149/) (2023)
- [Large language models surpass human experts in predicting neuroscience results](https://mlsanity.com/paper/2403.03230/) (2024)
- [SciAssess: Benchmarking LLM Proficiency in Scientific Literature Analysis](https://mlsanity.com/paper/2403.01976/) (2024)
- [PhysicsAssistant: An LLM-Powered Interactive Learning Robot for Physics Lab Investigations](https://mlsanity.com/paper/2403.18721/) (2024)

## Built on by (most cited first, 25 of 25)

- [Transforming Science with Large Language Models: A Survey on AI-assisted Scientific Discovery, Experimentation, Content Generation, and Evaluation](https://mlsanity.com/paper/2502.05151/) (2025)
- [Dissecting Logical Reasoning in LLMs: A Fine-Grained Evaluation and Supervision Study](https://mlsanity.com/paper/2506.04810/) (2025)
- [BioProBench: A Corpus and Benchmark for Biological Protocol Reasoning in Autonomous Science](https://mlsanity.com/paper/2505.07889/) (2025)
- [LLM-based HSE Compliance Assessment: Benchmark, Performance, and Advancements](https://mlsanity.com/paper/2505.22959/) (2025)
- [Measuring Harmfulness of Computer-Using Agents](https://mlsanity.com/paper/2508.00935/) (2025)
- [Artificial Intelligence in Spectroscopy: Advancing Chemistry from Prediction to Generation and Beyond](https://mlsanity.com/paper/2502.09897/) (2025)
- [Beyond Single-Value Metrics: Evaluating and Enhancing LLM Unlearning with Cognitive Diagnosis](https://mlsanity.com/paper/2502.13996/) (2025)
- [SAFER: Advancing Safety Alignment via Efficient Ex-Ante Reasoning](https://mlsanity.com/paper/2504.02725/) (2025)
- [Toward Autonomous Laboratory Safety Monitoring with Vision Language Models: Learning to See Hazards Through Scene Structure](https://mlsanity.com/paper/2602.00414/) (2026)
- [ForesightSafety Bench: A Frontier Risk Evaluation and Governance Framework towards Safe AI](https://mlsanity.com/paper/2602.14135/) (2026)
- [LABSHIELD: A Multimodal Benchmark for Safety-Critical Reasoning and Planning in Scientific Laboratories](https://mlsanity.com/paper/2603.11987/) (2026)
- [SENTINEL: A Multi-Level Formal Framework for Safety Evaluation of Foundation Model-based Embodied Agents](https://mlsanity.com/paper/2510.12985/) (2025)
- [LabGuard: Grounding Natural-Language Laboratory Rules into Runtime Guards for Embodied Laboratory Agents](https://mlsanity.com/paper/2606.31045/) (2026)
- [Mechanist: AI as a Scientific Instrument for Discovering the Mechanisms of Intelligence](https://mlsanity.com/paper/2608.12036/) (2026)
- [When Preferences Fail to Become Incentives: A Utility-Behavior Gap in Large Language Models](https://mlsanity.com/paper/2606.22974/) (2026)
- [SciHazard: A Benchmark for Measuring Scientific Safety Risks with Decomposed Harm Scoring](https://mlsanity.com/paper/2607.18665/) (2026)
- [Evaluating Large Language Model Performance on International Maritime Dangerous Goods Code Compliance](https://mlsanity.com/paper/2608.21036/) (2026)
- [Malicious Repurposing of Open Science Artefacts by Using Large Language Models](https://mlsanity.com/paper/2601.18998/) (2026)
- [Alignment Risks from Capability-Seeking RL Training](https://mlsanity.com/paper/2602.12124/) (2026)
- [Safe-SDL:Establishing Safety Boundaries and Control Mechanisms for AI-Driven Self-Driving Laboratories](https://mlsanity.com/paper/2602.15061/) (2026)
- [Beyond single-channel agentic benchmarking](https://mlsanity.com/paper/2602.18456/) (2026)
- [PseudoBench: Measuring How Agentic Auto-Research Fuels Pseudoscience](https://mlsanity.com/paper/2606.18060/) (2026)
- [Visual Compliance via Executable Safety Rule Entailment](https://mlsanity.com/paper/2609.18328/) (2026)
- [LabFactory: Building and Evaluating Executable AI Labs](https://mlsanity.com/paper/2609.28697/) (2026)
- [LabOPBench: A Multimodal Benchmark for Evidence-Grounded Laboratory Decision Making](https://mlsanity.com/paper/2610.19780/) (2026)

## Abstract (arXiv)

Artificial Intelligence (AI) is revolutionizing scientific research, yet its growing integration into laboratory environments presents critical safety challenges. Large language models (LLMs) and vision language models (VLMs) now assist in experiment design and procedural guidance, yet their "illusion of understanding" may lead researchers to overtrust unsafe outputs. Here we show that current models remain far from meeting the reliability needed for safe laboratory operation. We introduce LabSafety Bench, a comprehensive benchmark that evaluates models on hazard identification, risk assessment, and consequence prediction across 765 multiple-choice questions and 404 realistic lab scenarios, encompassing 3,128 open-ended tasks. Evaluations on 19 advanced LLMs and VLMs show that no model evaluated on hazard identification surpasses 70% accuracy. While proprietary models perform well on structured assessments, they do not show a clear advantage in open-ended reasoning. These results underscore the urgent need for specialized safety evaluation frameworks before deploying AI systems in real laboratory settings.