from 01.01.2013 until now
Russian Federation
Russian Federation
UDC 614.8
Modern artificial intelligence technologies, particularly large language models, demonstrate significant potential for transforming approaches to emergency safety. The purpose of this article is to provide an analytical review and assess the potential for using large language models in emergency safety. Based on this review, the article systematizes and analyzes in detail promising areas of large language models application, including: automation of emergency message and call processing; creation of intelligent chatbots and virtual assistants for the public and security professionals; planning and decision support; social media data analysis to improve situational awareness; working with multimodal data; development of highly specialized models for specific subject areas (forest fires, evacuation, etc.); integration with expert systems and knowledge bases; and personnel training. Particular attention is paid to analyzing the advantages and critical limitations of these technologies, such as the problem of large language models «hallucinations», and ways to minimize them. The importance of adapting models to national specifics, including language features, regulatory frameworks, and local risks, is emphasized. The significance of this study lies in its ability to provide a comprehensive understanding of the current level of development and future trajectories of large language models integration into emergency safety systems. The study demonstrates that the proper implementation of language models can significantly improve response times, the validity of management decisions, and the effectiveness of interagency cooperation, serving as a powerful tool for intellectual support in decision-making in the field of security.
large language models, emergency safety, intelligent decision support, natural language processing, transformer
1. Raza M. et al. Industrial applications of large language models // Scientific Reports. 2025. Vol. 15. № 1. P. 13755. DOI:https://doi.org/10.1038/s41598-025-98483-1.
2. Filimonov V.Yu. Bol'shie yazykovye modeli i ih rol' v sovremennyh nauchnyh otkrytiyah // Filosofskie problemy informacionnyh tekhnologij i kiberprostranstva. 2024. № 1 (25). S. 42–57. DOI:https://doi.org/10.17726/philIT.2024.1.3. EDN SFTWKC.
3. Otal H.T., Stern E., Canbaz M.A. LLM-assisted crisis management: Building advanced LLM platforms for effective emergency response and public collaboration // 2024 IEEE Conference on Artificial Intelligence (CAI). IEEE, 2024. P. 851–859. DOI:https://doi.org/10.1109/CAI59869.2024.00159.
4. From text to transformation: A comprehensive review of large language models\' versatility / P. Kaur [et al.] // arXiv preprint arXiv:2402.16142. 2024. DOI:https://doi.org/10.48550/arXiv.2402.16142.
5. Panina O.A., Yurin D.A., Suhov I.Yu. Riski bezopasnosti pri ispol'zovanii bol'shih yazykovyh modelej (LLM) v sistemah // Informatizaciya i svyaz'. 2024. № 6. S. 65–72. DOI:https://doi.org/10.34219/2078-8320-2024-15-6-65-72. EDN NCCJVT.
6. Chandra A., Chakraborty A. Exploring the role of large language models in radiation emergency response // Journal of Radiological Protection. 2024. Vol. 44. № 1. P. 011510. DOI:https://doi.org/10.1088/1361-6498/ad270c.
7. Jiang Y. The Applications of Large Language Models in Emergency Management // 2024 IEEE 6th Advanced Information Management, Communicates, Electronic and Automation Control Conference (IMCEC). IEEE, 2024. Vol. 6. P. 199–202. DOI:https://doi.org/10.1109/IMCEC59810.2024.10575031.
8. The role of large language models for decision support in fire safety planning / D. Durmus [et al.] // PROCEEDINGS OF THE... ISARC. 2024. P. 339–346. DOI:https://doi.org/10.22260/ISARC2024/0045.
9. Chuchupal V.Ya. Nejrosetevye modeli yazyka dlya sistem raspoznavaniya rechi // Rechevye tekhnologii. 2020. № 1-2. S. 27–47. DOI:https://doi.org/10.58633/2305-8129_2020_1-2_27. EDN TOXWRW.
10. Attention is all you need / A. Vaswani [et al.] // Advances in neural information processing systems. 2017. Vol. 30. DOI:https://doi.org/10.48550/arXiv.1706.03762.
11. SafeMate: A Modular RAG-Based Agent for Context-Aware Emergency Guidance / J. Jiao [et al.] // arXiv preprint arXiv:2505.02306. 2025. DOI:https://doi.org/10.48550/arXiv.2505.02306.
12. LLM multi-agent systems: Challenges and open problems / S. Han [et al.] // arXiv preprint arXiv:2402.03578. 2024. DOI:https://doi.org/10.48550/arXiv.2402.03578.
13. Gupta S., Chen Y.C., Tsai C. Utilizing large language models in tribal emergency management // Companion Proceedings of the 29th International Conference on Intelligent User Interfaces. 2024. P. 1–6. DOI:https://doi.org/10.1145/3640544.3645219.
14. California launches new AI-powered chatbot that provides wildfire resources in 70 languages. URL: https://www.gov.ca.gov/2025/05/09/california-launches-new-ai-powered-chatbot-that-provides-wildfire-resources-in-70-languages/?utm_source=chatgpt.com (data obrashcheniya: 03.08.2025)
15. Large language models in fire engineering: An examination of technical questions against domain knowledge / H. Hostetter [et al.] // arXiv preprint arXiv:2403.04795. 2024. DOI:https://doi.org/10.48550/arXiv.2403.04795.
16. The ERMES chatbot: A conversational communication tool for improved emergency management and disaster risk reduction / A. Urbanelli [et al.] // International Journal of Disaster Risk Reduction. 2024. Vol. 112. P. 104792. DOI:https://doi.org/10.1016/j.ijdrr.2024.104792.
17. PathGPT: Leveraging Large Language Models for Personalized Route Generation / S.C. Marcelyn [et al.] // arXiv preprint arXiv:2504.05846. 2025. DOI:https://doi.org/10.48550/arXiv.2504.05846.
18. Maksimov A.V., Matveev A.V. Perspektivy ispol'zovaniya kollektivnyh znanij pri reagirovanii na chrezvychajnye situacii // Nauch.-analit. zhur. «Vestnik S.-Peterb. un-ta GPS MCHS Rossii». 2019. № 4. S. 89–97. EDN QPBTLA.
19. Metodika analiza dannyh o chrezvychajnyh situaciyah v social'nyh setyah / A.V. Vostryh [i dr.] // Sovremennye naukoemkie tekhnologii. 2023. № 6. S. 81–88. DOI:https://doi.org/10.17513/snt.39635. EDN ZCOPNS.
20. Crisissense-llm: Instruction fine-tuned large language model for multi-label social media text classification in disaster informatics / K. Yin [et al.] // arXiv preprint arXiv:2406.15477. 2024. DOI:https://doi.org/10.48550/arXiv.2406.15477.
21. Lamsal R., Read M.R., Karunasekera S. CrisisTransformers: Pre-trained language models and sentence encoders for crisis-related social media texts // Knowledge-Based Systems. 2024. Vol. 296. P. 111916. DOI:https://doi.org/10.1016/j.knosys.2024.111916.
22. Fireexpert: Fire event identification and assessment leveraging cross-domain knowledge and large language model / G. Luo [et al.] // IEEE Transactions on Mobile Computing. 2025. DOI:https://doi.org/10.1109/TMC.2025.3528413.
23. A real-time mapping method for knowledge graph-driven large language models: a focus on indoor fire evacuations / T. Wang [et al.] // International Journal of Digital Earth. 2025. Vol. 18. № 1. P. 2468407. DOI:https://doi.org/10.1080/17538947.2025.2468407.
24. Wildfiregpt: Tailored large language model for wildfire analysis / Y. Xie [et al.] // arXiv preprint arXiv:2402.07877. 2024. DOI:https://doi.org/10.48550/arXiv.2402.07877.
25. Using large language models to extract core injury information from emergency department notes / D.H. Choi [et al.] // Journal of Korean Medical Science. 2024. Vol. 39. № 46. DOI:https://doi.org/10.3346/jkms.2024.39.e291.
26. Sim911: Towards effective and equitable 9-1-1 dispatcher training with an llm-enabled simulation / Z. Chen [et al.] // Proceedings of the AAAI Conference on Artificial Intelligence. 2025. Vol. 39. № 27. P. 27896–27904. DOI:https://doi.org/10.1609/aaai.v39i27.35006.
