Analysis of Internet Texts with Signs of Information Threats Using Large Language Models
УДК 81’3, ББК 81.1
Abstract
The article aims to verify the hypothesis about the ability of ChatGPT and DeepSeek language models to analyze natural language texts in order to identify language means that may indicate an information security threat associated with extremist activities. The methodological basis of the research is a fundamentally new approach to the study of Internet texts through digital technologies from the standpoint of information security, taking into account the legal aspects of information threats research, including the threat of extremism. The analyzed material consists of 100 Internet texts in the Russian language, publicly available on the social networks VKontakte and Odnoklassniki, selected by random sampling from these networks’ open Christian communities. The semantic analysis of the selected texts allowed us to divide them into groups reflecting their subjective sentiment: neutral, approving, insulting, condemning and containing a strong negative assessment. The last three groups fall into the risk of containing an information threat. The research showed that both ChatGPT and DeepSeek systems adequately cope with the task of identifying language expressions that require further analysis for possible threats to information security. It is important to note that the responses given by Large Language Models can only be used as an additional instrument to support an expert assessment of the text for signs of information threats. This instrument helps to look at the analyzed text more objectively, separately from the personal attitudes of a human expert. However, only a human expert can make a final conclusion about the existence of an information threat.
Downloads
References
Грамота.ру, 2000–2025. URL: https://gramota.ru/poisk?query=тварь&mode=slovari&dicts[]=42
Грушихина В. А. Судебная психолого-лингвистическая экспертиза текста (на примере дела об экстремизме) / Деятельность правоохранительных органов в современных условиях. Сборник материалов XXI международной научно-практической конференции. – Иркутск, 2016. – С. 12–15.
Доктрина информационной безопасности Российской Федерации. Утверждена Указом Президента Российской Федерации от 5 декабря 2016 г. №646. URL: https://www.mid.ru/ru/foreign_policy/official_documents/1539546/.
КоАП РФ Статья 13.53. Поиск заведомо экстремистских материалов и получение доступа к ним, в том числе с использованием программно-аппаратных средств доступа к информационным ресурсам, информационно-телекоммуникационным сетям, доступ к которым ограничен. URL: https://www.consultant.ru/document/cons_doc_LAW_34661/d698a1c3214cb529c29991c00154e985d9d7f607/
Комов С. А. Термины и определения в области информационной безопасности. Москва, 2009.
Королькова А. В., Алексеева М. И. Лингвистическая экспертиза текстов, публично оправдывающих терроризм и/или экстремизм / Филологические науки. Вопросы теории и практики. – 2023. – Т. 16. – № 8. – С. 2339–2343.
Куршакова Т. Д., Рядчикова Е. Н., Схаляхова С. Ш. Языковые маркеры экстремизма, фашизма, нацизма: из практики судебных лингвистических экспертиз / Казанская наука. – 2023. – № 6. – С. 27–33.
Овсянникова М.А. Анализ тональности текста посредством инструмента искусственного интеллекта / Три «Л» в парадигме современного гуманитарного знания: Лингвистика, Литературоведение, Лингводидактика. Москва, 2025. С. 33–38.
Юлдашев Ж. Б. К определению сущности международной информационной безопасности / Коммуникология: электронный научный журнал. – 2022. – Т. 7. – № 4. – С. 115–125.
Bing Liu. Sentiment Analysis and Subjectivity / Handbook of Natural Language Processing (англ.) / под ред. N. Indurkhya и F. J. Damerau. 2010. URL: https://www.cs.uic.edu/~liub/FBS/NLP-handbook-sentiment-analysis.pdf
Fullstack Academy, LLC. What is Information Security? 2025. URL: https://www.fullstackacademy.com/blog/what-is-information-security
Joshi Ch., Singh U.K. Information security risks management framework – A step towards mitigating security risks in university network / Journal of Information Security and Applications. – 2017. – No. 35. – Pp. 128–137. doi:10.1016/j.jisa.2017.06.006.
TechTarget Contributor, 2024, Top 10 types of information security threats for IT teams, URL: https://www.techtarget.com/searchsecurity/feature/Top-10-types-of-information-security-threats-for-IT-teams.
Copyright (c) 2025 Светлана Осокина, Павел Осокин
This work is licensed under a Creative Commons Attribution 4.0 International License.
The authors, which are published in this journal, agree to the following conditions:
1. Authors retain the copyright to the work and transfer to the journal the right of the first publication along with the work, at the same time licensing it under the terms of the Creative Commons Attribution License, which allows others to distribute this work with the obligatory indication of the authorship of this work and a link to the original publication in this journal .
2. The authors retain the right to enter into separate, additional contractual agreements for the non-exclusive distribution of the version of the work published by this journal (for example, to place it in the university depository or to publish it in a book), with reference to the original publication in this journal.
3. Authors are allowed to post their work on the Internet (for example, in a university repository or on their personal website) before and during the review process of this journal, as this may lead to a productive discussion, as well as more links to this published work (See The Effect of Open Access).
