Comparison of Capabilities of Modern LLM Models in Analyzing and Developing National Security Strategies

Tomasz Zawadzki

doi:10.35467/cal/214578

Akademii Sztuki Wojennej

PL EN

Strona główna / Publikacje / Comparison of Capabilities of Modern LLM Models in Analyzing and Developing National Security Strategies

Zgłoś uwagę

Artykuł

Pobierz plik Pobierz BibTeX

Tytuł

Comparison of Capabilities of Modern LLM Models in Analyzing and Developing National Security Strategies

Autorzy

Tomasz Zawadzki ^{[ 1 ][ 6.3 ][ P ]}

^{[ 1 ]} Faculty od Mechatronics, Warsaw University of Technology, Poland | ^{[ P ]} pracownik

Dyscyplina naukowa (Ustawa 2.0)

[6.3] Nauki o bezpieczeństwie

Rok publikacji

2025

Opublikowano w

Cybersecurity and Law

Rocznik: 2025 | Numer: nr 1 (13)

Typ artykułu

artykuł naukowy

Język publikacji

angielski

Słowa kluczowe

EN

Cybersecurity
Large Language Models (LLMs)
Modelling
National security
Technology

PL

Bezpieczeństwo narodowe
Cyberbezpieczeństwo
Modelowanie
Technologia

Streszczenie

EN Objectives: The aim of the research described in this article was to assess the feasibility of using offline Large Language Models (LLMs) to analyze contemporary cybersecurity strategies of individual countries, identify their strengths and weaknesses, and subsequently prepare recommendations for designing national strategies. Methods: The research methodology adopted by the author first involved setting up the research environment, which included configuring hardware and software to enable efficient work with LLMs, followed by interacting with the models through appropriate queries and analyzing their responses. Depending on the answers provided, follow-up questions were asked for clarification. Three of the most popular offline models were selected for interaction: LLaMA, QwQ, and DeepSeek-R1, each receiving the same set of questions. Results: The outcome of these queries was a comparative analysis of the cybersecurity strategies of the USA, Germany, and Poland, as well as the strategy of the War Studies University, highlighting identified differences and similarities, and then preparing recommendations for a hypothetical Polish cybersecurity strategy for the years 2025–2035. Conclusions: Based on the analysis of responses generated by individual models, it can be concluded that each of the examined models is capable of bringing unique value to the strategy development process, and their combination can significantly streamline the creation of a comprehensive strategic document for Poland, considering that: - LLAMA 3.3 performs well as a document structure generator - QwQ is the most advanced in strategic planning, budgeting, and risk analysis - DeepSeek-R1 demonstrates high flexibility and the ability to interpret context, even with limited knowledge

Data udostępnienia online

21.11.2025

Strony (od-do)

206 - 218

DOI

10.35467/cal/214578

URL

https://czasopismocybersecurityandlaw.pl/Comparison-of-Capabilities-of-Modern-LLM-Models-in-Analyzing-and-Developing-National,214578,0,1.html

Uwagi

Bibliografia, netografia na stronach 217-218.

Typ licencji

CC BY-SA (uznanie autorstwa - na tych samych warunkach)