Elamur Research⌕
☰
home / papers / 2026 / Can Large Language Models Effectively Support Decision-Making in Sudden Emergencies?

Can Large Language Models Effectively Support Decision-Making in Sudden Emergencies?

Paper · 2026 · Mengna Zhu, Jibing Wu, Lihua Liu et al.

Summary

Large language models are increasingly applied to emergency management, but most prior work stops at understanding events, for example through extraction or summarization. The step where an actual response decision is made has received little attention, largely because no suitable data existed. This paper defines emergency decision-making as reasoning over an event description together with applicable regulations, producing both an answer and a traceable justification. The authors release EDM-Bench, a Chinese-language benchmark built from real incidents. It covers verification questions, option-selection questions, free-form answers and full response reports. They also release EDM-R², a repository of regulations and contingency plans split into structured, retrievable items. On top of these resources they propose R³V-EDM, a pipeline that retrieves relevant rules, constrains reasoning with them, builds an evidence chain, and then verifies and refines the output. Benchmarking of current LLMs shows the task is genuinely difficult, and rule-grounded reasoning makes decisions more reliable and easier to interpret. This matters for safety-critical public governance, where invented regulations or wrong authority assignments are unacceptable.

Key points

  • EDM-Bench has 1,179 instances in two subsets, EDM-QA and EDM-Report. It covers five formats: TF, MCQ, MAQ, SAG and RG.
  • The data span four emergency categories and 11 event types: natural disasters, accidents, public health and social security.
  • EDM-R² holds 3,406 structured rules parsed from 38 official documents, about 3.5 per event type. Each rule is itemized by trigger condition, responsible authority and required action.
  • QA candidates were drafted by an LLM (Qwen3-Max appears in the pipeline) and then reviewed by human annotators. Reports were written manually from real reports and expert templates. Eight NLP graduate students did the annotation.
  • The authors list three limitations:
    • rule coverage is restricted to the included event types;
    • the dataset is small;
    • models sometimes reach correct answers by "lucky guessing" with flawed reasoning, which calls for automated evaluation of reasoning paths.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

 Institutions

🎓Tsinghua UniversityBeijing · China415 works · 94,028 citations🎓National University of Defense TechnologyChangsha · China20 works · 1,339 citations🎓University of Electronic Science and Technology of ChinaChengdu · China38 works · 2,088 citations

 Fields, topics and concepts

📚 Shelves 8The field of science this work belongs to (OpenAlex fields).Computer Science · 18,614 worksArtificial Intelligence · 17,965 worksArts and Humanities · 17,450 worksLanguage and Linguistics · 17,296 worksSocial Sciences · 1,954 worksPolitical Science and International Relations · 206 worksBusiness, Management and Accounting · 40 worksOrganizational Behavior and Human Resource Management · 15 works

🧭 Topics 2Research subjects and languages the work is about.Language models & pretraining · 3,894 worksReasoning · 1,685 works

Cite

Cite as Mengna Zhu, Jibing Wu, Lihua Liu, Yuran Gong, Yang Hao, Fu Yachao, Mao Wang, Lei Hou and Juanzi Li. 2026. Can Large Language Models Effectively Support Decision-Making in Sudden Emergencies?. Findings of the Association for Computational Linguistics: ACL 2026.

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/2026.findings-acl.1820

@inproceedings{zhu2026can,
  title = {{Can Large Language Models Effectively Support Decision-Making in Sudden Emergencies?}},
  author = {Mengna Zhu, Jibing Wu, Lihua Liu, Yuran Gong, Yang Hao, Fu Yachao, Mao Wang, Lei Hou and Juanzi Li},
  booktitle = {Findings of the Association for Computational Linguistics: ACL 2026},
  year = {2026},
  doi = {10.18653/v1/2026.findings-acl.1820},
  url = {https://doi.org/10.18653/v1/2026.findings-acl.1820}
}

References

33 references, 8 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Anthropic. 2025. Introducing claude sonnet 4.5. https://www.anthropic.com/news/claude-sonnet-4-5. ↗
  2. Yuntao Bai, Saurav Kadavath, Sandipan Kundu, Amanda Askell, Jackson Kernion, Andy Jones, Anna Chen, Anna Goldie, Azalia Mirhoseini, Cameron McKinnon, and 1 others. 2022. Constitutional ai: Harmlessness from ai feedback. arXiv preprint arXiv:2212.08073.
  3. Yushi Bai, Shangqing Tu, Jiajie Zhang, Hao Peng, Xiaozhi Wang, Xin Lv, Shulin Cao, Jiazheng Xu, Lei Hou, Yuxiao Dong, and 1 others. 2025. Longbench v2: Towards deeper understanding and reasoning on realistic long-context multitasks. In Proceedings of the 63rd Annual Meeting of the Association for Co…
  4. Gabriela Pajtinková Bartáková, Ahmad Almadhor, Amna Qayyum, Kainaat Abeer, Aman Durrani, and 1 others. 2025. Evaluating the capacity and limitations of generative ai in financial decision making. Computer Standards & Interfaces, 93:103965. find
  5. Minze Chen, Zhenxiang Tao, Weitong Tang, Tingxin Qin, Rui Yang, and Chunli Zhu. 2024. Enhancing emergency decision-making with knowledge graphs and large language models. International Journal of Disaster Risk Reduction, 113:104804. find
  6. Shehzaad Dhuliawala, Mojtaba Komeili, Jing Xu, Roberta Raileanu, Xian Li, Asli Celikyilmaz, and Jason Weston. 2024. Chain-of-verification reduces hallucination in large language models. In Findings of the association for computational linguistics: ACL 2024, pages 3563–3578.
  7. Hossein Rajaby Faghihi, Bashar Alhafni, Ke Zhang, Shihao Ran, Joel Tetreault, and Alejandro Jaimes. 2022. Crisisltlsum: A benchmark for local crisis event timeline extraction and summarization. arXiv preprint arXiv:2210.14190. find
  8. Demian Gholipour Ghalandari, Chris Hokamp, John Glover, Georgiana Ifrim, and 1 others. 2020. A large-scale multi-document summarization dataset from the wikipedia current events portal. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 1302–1308. find
  9. Team GLM, Aohan Zeng, Bin Xu, Bowen Wang, Chenhui Zhang, Da Yin, Diego Rojas, Guanyu Feng, Hanlin Zhao, Hanyu Lai, Hao Yu, Hongning Wang, Jiadai Sun, Jiajie Zhang, Jiale Cheng, Jiayi Gui, Jie Tang, Jing Zhang, Juanzi Li, and 37 others. 2024. Chatglm: A family of large language models from glm-130b t…
  10. Ollama. 2024. Llama 3.1. https://ollama.com/library/ll ama3.1. ↗
  11. Yun He, Wenzhe Li, Hejia Zhang, Songlin Li, Karishma Mandyam, Sopan Khosla, Yuanhao Xiong, Nanshu Wang, Xiaoliang Peng, Beibin Li, and 1 others. 2025. Advancedif: Rubric-based benchmarking and reinforcement learning for advancing llm instruction following. arXiv preprint arXiv:2511.10507.
  12. Ollama. 2025. Llama 3.3. https://ollama.com/library/ll ama3.3. ↗
  13. Zhuowei Hu, Huili Gong, and Liying Zhu. 2007. Fast flooding information extraction in emergency response of flood disaster. In Proceedings of the ISPRS workshop on updating geo-spatial databases with imagery and the 5th ISPRS workshop on DMGISs, Urumchi, pages 28–29. find
  14. Ahrii Kim. 2025. Rubric-mqm: Span-level llm-as-judge in machine translation for high-end models. In Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 6: Industry Track), pages 147–165. find
  15. Tomek Korbak, Mikita Balesni, Elizabeth Barnes, Yoshua Bengio, Joe Benton, Joseph Bloom, Mark Chen, Alan Cooney, Allan Dafoe, Anca Dragan, and 1 others. 2025. Chain of thought monitorability: A new and fragile opportunity for ai safety. arXiv preprint arXiv:2507.11473. find
  16. Jia Li, Zichun Zhou, Han Lyu, and Zhenchang Wang. 2025. Large language models-powered clinical decision support: enhancing or replacing human expertise? find
  17. Jiaqi Li, Mengmeng Wang, Zilong Zheng, and Muhan Zhang. 2024. Loogle: Can long-context language models understand long contexts? In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 16304–16333.
  18. Aixin Liu, Bei Feng, Bing Xue, Bingxuan Wang, Bochao Wu, Chengda Lu, Chenggang Zhao, Chengqi Deng, Chenyu Zhang, Chong Ruan, and 1 others. 2024. Deepseek-v3 technical report. arXiv preprint arXiv:2412.19437. find
  19. Jiaheng Liu, Dawei Zhu, Zhiqi Bai, Yancheng He, Huanxuan Liao, Haoran Que, Zekun Wang, Chenchen Zhang, Ge Zhang, Jiebin Zhang, and 1 others. 2025. A comprehensive survey on long context language modeling. arXiv preprint arXiv:2503.17407.
  20. Aman Madaan, Niket Tandon, Prakhar Gupta, Skyler Hallinan, Luyu Gao, Sarah Wiegreffe, Uri Alon, Nouha Dziri, Shrimai Prabhumoye, Yiming Yang, and 1 others. 2023. Self-refine: Iterative refinement with self-feedback. Advances in neural information processing systems, 36:46534–46594. find
  21. Heng Mu, Peng Wu, and Wenyi Su. 2024. Construction of knowledge graph for emergency resources. International Journal of Intelligent Systems, 2024(1):6668559. find
  22. OpenAI. 2025. Update to gpt-5 system card: Gpt-5.2. find
  23. Aditya Pathak, Rachit Gandhi, Vaibhav Uttam, Yashwanth Nakka, Aaryan Raj Jindal, Pratyush Ghosh, Arnav Ramamoorthy, Shreyash Verma, Aditya Mittal, Aashna Ased, and 1 others. 2025. Rubric is all you need: Enhancing llm-based code evaluation with question-specific rubrics. arXiv preprint arXiv:2503.23… — [arXiv:2503.23989](https://arxiv.org/abs/2503.23989) find
  24. Jayr Pereira, Andre Assumpcao, Julio Trecenti, Luiz Airosa, Caio Lente, Jhonatan Cléto, Guilherme Dobins, Rodrigo Nogueira, Luis Mitchell, and Roberto Lotufo. 2025. Inacia: Integrating large language models in brazilian audit courts: Opportunities and challenges. Digital Government: Research and Pra… find
  25. Eric A Posner and Shivam Saran. 2025. Judge ai: Assessing large language models in judicial decision-making. University of Chicago Coase-Sandor Institute for Law & Economics Research Paper, (2503). find
  26. Qwen, :, An Yang, Baosong Yang, Beichen Zhang, Binyuan Hui, Bo Zheng, Bowen Yu, Chengyuan Li, Dayiheng Liu, Fei Huang, Haoran Wei, Huan Lin, Jian Yang, Jianhong Tu, Jianwei Zhang, Jianxin Yang, Jiaxi Yang, Jingren Zhou, and 25 others. 2025. Qwen2.5 technical report. Preprint, arXiv:2412.15115. find
  27. Nils Reimers and Iryna Gurevych. 2019. Sentence-bert: Sentence embeddings using siamese bert-networks. arXiv preprint arXiv:1908.10084.
  28. Sarah Sandmann, Stefan Hegselmann, Michael Fujarski, Lucas Bickmann, Benjamin Wild, Roland Eils, and Julian Varghese. 2025. Benchmark evaluation of deepseek large language models in clinical decision-making. Nature Medicine, pages 1–1. find
  29. Wenxin Su, Linyan Chen, and Xin Gao. 2022. Emergency decision making: A literature review and future directions. Sustainability, 14(17):10925. find
  30. Jianzhuo Yan, Lihong Chen, Yongchuan Yu, Hongxia Xu, Qingcai Gao, Kunpeng Cao, and Jianhui Chen. 2022. Emergeventmine: End-to-end chinese emergency event extraction using a deep adversarial network. ISPRS International Journal of Geo-Information, 11(6):345. find
  31. Qingyuan Zhang, Yuxi Wang, Cancan Hua, Yulin Huang, and Ning Lyu. 2025. Knowledge-augmented large language model agents for explainable financial decision-making. arXiv preprint arXiv:2512.09440. — [arXiv:2512.09440](https://arxiv.org/abs/2512.09440) find
  32. Zi-Xin Zhang, Liang Wang, and Ying-Ming Wang. 2018. An emergency decision making method based on prospect theory for different emergency situations. International Journal of Disaster Risk Science, 9(3):407–420. find
  33. Mengna Zhu, Kaisheng Zeng, Mao Wang, Kaiming Xiao, Lei Hou, Hongbin Huang, and Juanzi Li. 2025. Eventsum: A large-scale event-centric summarization dataset for chinese multi-news documents. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 39, pages 26138–26147. find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/zhu-2026-can-large-language-models-effectively/