NoobGPT: LLMs e a geração de malwares indetectáveis
Resumo
Este artigo explora a capacidade do modelo de linguagem ChatGPT gerar malwares por meio de instruções conhecidas como jailbreaks, amplamente divulgados na internet. A pesquisa simula o uso por um usuário leigo, sem conhecimento técnico em cibersegurança, e avalia os códigos maliciosos gerados quanto à sua funcionalidade e detecção por softwares de segurança. Os testes mostraram que é possível obter diferentes tipos de malware, com a maioria sendo indetectável já nas primeiras interações. Observou-se também que os mecanismos de segurança, embora presentes, podem ser contornados. Os resultados levantam preocupações sobre o uso indevido de LLMs e os limites atuais das ferramentas de proteção digital.Referências
Gupta, M., Akiri, C., Aryal, K., Parker, E., e Praharaj, L. (2023). From ChatGPT to ThreatGPT: Impact of generative AI in cybersecurity and privacy. IEEE Access, 11, pp. 80218–80245.
Liu, Y., Deng, G., Xu, Z., Li, Y., Zheng, Y., Zhang, Y., Zhao, L., Zhang, T., Wang, K., & Liu, Y. (2024). Jailbreaking ChatGPT via prompt engineering: An empirical study (arXiv:2305.13860). arXiv.
Madani, P. (2023). Metamorphic malware evolution: The potential and peril of large language models. In: 5th IEEE International Conference on Trust, Privacy and Security in Intelligent Systems and Applications (TPS-ISA), pp. 74–81.
OpenAI. (2025). Using GPTs on our Free Tier (FAQ). [link].
OpenAI Community. (2025). Interaction limits for ChatGPT-4. [link].
Pa, Y. M., Tanizaki, S., Kou, T., Van Eeten, M., Yoshioka, K., e Matsumoto, T. (2023). An attacker’s dream? Exploring the capabilities of ChatGPT for developing malware. In: Proceedings of the 16th Cyber Security Experimentation and Test Workshop.
Stanford University. (2024). The 2024 AI Index Report. [link].
Tahir, R. (2018). A study on malware and malware detection techniques. International Journal of Education and Management Engineering, 8(2), p. 20. Modern Education and Computer Science Press.
Xu, Z., Liu, Y., Deng, G., Li, Y., & Picek, S. (2024). A comprehensive study of jailbreak attack versus defense for large language models. In: Findings of the Association for Computational Linguistics: ACL 2024 (pp. 7432–7449).
Yamin, M. M., Hashmi, E., e Katt, B. (2024). Combining uncensored and censored LLMs for ransomware generation. In: International Conference on Web Information Systems Engineering, pp. 189–202. Springer.
Yong, Z. X., Menghini, C., & Bach, S. (2023). Low-resource languages jailbreak GPT-4. In: Socially Responsible Language Modelling Research.
Yong Wong, M., Landen, M., Antonakakis, M., Blough, D. M., Redmiles, E. M., e Ahamad, M. (2021). An inside look into the practice of malware analysis. In Proceedings of the 2021 ACM SIGSAC Conference on Computer and Communications Security, pp. 3053–3069.
Zhu, K. (2024). Ranked: The most popular generative AI tools in 2024. Visual Capitalist. [link].
Liu, Y., Deng, G., Xu, Z., Li, Y., Zheng, Y., Zhang, Y., Zhao, L., Zhang, T., Wang, K., & Liu, Y. (2024). Jailbreaking ChatGPT via prompt engineering: An empirical study (arXiv:2305.13860). arXiv.
Madani, P. (2023). Metamorphic malware evolution: The potential and peril of large language models. In: 5th IEEE International Conference on Trust, Privacy and Security in Intelligent Systems and Applications (TPS-ISA), pp. 74–81.
OpenAI. (2025). Using GPTs on our Free Tier (FAQ). [link].
OpenAI Community. (2025). Interaction limits for ChatGPT-4. [link].
Pa, Y. M., Tanizaki, S., Kou, T., Van Eeten, M., Yoshioka, K., e Matsumoto, T. (2023). An attacker’s dream? Exploring the capabilities of ChatGPT for developing malware. In: Proceedings of the 16th Cyber Security Experimentation and Test Workshop.
Stanford University. (2024). The 2024 AI Index Report. [link].
Tahir, R. (2018). A study on malware and malware detection techniques. International Journal of Education and Management Engineering, 8(2), p. 20. Modern Education and Computer Science Press.
Xu, Z., Liu, Y., Deng, G., Li, Y., & Picek, S. (2024). A comprehensive study of jailbreak attack versus defense for large language models. In: Findings of the Association for Computational Linguistics: ACL 2024 (pp. 7432–7449).
Yamin, M. M., Hashmi, E., e Katt, B. (2024). Combining uncensored and censored LLMs for ransomware generation. In: International Conference on Web Information Systems Engineering, pp. 189–202. Springer.
Yong, Z. X., Menghini, C., & Bach, S. (2023). Low-resource languages jailbreak GPT-4. In: Socially Responsible Language Modelling Research.
Yong Wong, M., Landen, M., Antonakakis, M., Blough, D. M., Redmiles, E. M., e Ahamad, M. (2021). An inside look into the practice of malware analysis. In Proceedings of the 2021 ACM SIGSAC Conference on Computer and Communications Security, pp. 3053–3069.
Zhu, K. (2024). Ranked: The most popular generative AI tools in 2024. Visual Capitalist. [link].
Publicado
17/09/2025
Como Citar
CARVALHO, Gustavo Lofrese; LADEIRA, Ricardo de la Rocha; LIMA, Gabriel Eduardo.
NoobGPT: LLMs e a geração de malwares indetectáveis. In: WORKSHOP DE SISTEMAS DE INFORMAÇÃO (WSIS), 16. , 2025, Rio Paranaíba/MG.
Anais [...].
Porto Alegre: Sociedade Brasileira de Computação,
2025
.
p. 220-225.
DOI: https://doi.org/10.5753/wsis.2025.15064.
