Using generative AI tools in teaching students programming

   Introduction. Generative neural network tools are currently used for text processing, thesis writing, problem solving, and more. Their capabilities also allow for code generation, interpretation, and correction. Uncontrolled use of such services by students leads to a lower understanding of programming fundamentals and an inability to develop code optimization skills. In a context where it is impossible to limit the use of generative neural network tools, teaching methods must be modified to incorporate the fundamentals of prompt engineering.

   Purpose setting. This article examines the following AI services for teaching programming: ChatGPT, DeepSeek, Cursor, OpenAI Codex, Qwen, Gemini, Claude, and Perplexity.

   The purpose of this study is to justify the need to introduce prompt engineering into undergraduate IT curricula.

   Based on an analysis of scientific literature and the authors’ own teaching experience, the authors highlight the specific features of using generative neural network services in programming. A statistical study of students’ preferences and experiences with them was conducted.

   Methodology and methods of the study. The study utilized methods of systemic and comparative analysis, questionnaires, and data aggregation and statistical processing. The most common generative neural network services used by students in their programming studies are examined, and their advantages and disadvantages are discussed. A survey of students majoring in Applied Computer Science was also conducted to determine the frequency and specifics of their use of generative neural network assistants in educational activities. Empirical data was obtained through summarizing, grouping, and calculating relative indicators.

   Results. The paper highlights the key functions of generative neural network assistants most frequently used by students in their educational activities. These include creating program code from scratch, expanding and correcting existing programs, finding and fixing errors, clarifying algorithmic logic, mastering new technologies and libraries, and preparing reports. The analysis revealed that ChatGPT and DeepSeek are the most popular services, while Cursor and OpenAI Codex are typically used as development aids, and Claude and Perplexity are primarily used for information retrieval and clarification of complex concepts. Furthermore, the paper identifies the key pedagogical risks associated with the use of generative neural network assistants, including decreased student independence, superficial completion of learning tasks, and an increased risk of academic plagiarism.

   Conclusion. The paper acknowledges the need to incorporate prompt engineering into undergraduate IT curricula.

1. Bazhenova I. V. Features of the methodology of teaching programming based on projective-recursive strategy and cognitive technologies. Pedagogicheskoye obrazovaniye v Rossii, 2015, no. 3, pp. 52–57. (In Russ.).

2. Alferyeva-Termsikos V. B. Prompt engineering as a strategy for developing students’ information culture. Mezhdunarodnyy zhurnal gumanitarnykh i yestestvennykh nauk, 2024, no. 9–1 (96), pp. 10–15. (In Russ.).

3. Vukojicic M., Krstic J. ChatGPT in programming education: ChatGPT as a programming assistant. Inspired Teachers Voice, 2023, vol. 2023. no. 1. pp. 7–13.

4. Chen E. et al. GPTutor: a ChatGPT-powered programming tool for code explanation. International conference on artificial intelligence in education, Cham: Springer Nature Switzerland, 2023, pp. 321–327.

5. Allal L. B., Li R., Kocetkov D., Mou C., Akiki C., Ferrandis C. M., Muennighoff N., Mishra M., Gu A., Dey M. et al. SantaCoder: Don’t Reach for the Stars. Proceedings of the 17^th International Conference on Software Engineering for Adaptive and Self-Managing Systems, 2023, pp.1–12.

6. Barke S., James M. B., Polikarpova N. Grounded copilot: How programmers interact with code-generating models. Proceedings of the ACM on Programming Languages, 2023, vol. 7, no. OOPSLA1, pp. 85–111.

7. Ziegler, A., Kalliamvakou, E., Li, X., Rice, A., Rifkin, D., Simister, S., Sittampalam, G., Nenadic, A. Productivity, Efficiency, and Safety in AI-Assisted Software Development: An Empirical Study of GitHub Copilot. arXiv preprint arXiv:2206.15075, 2022, pp. 1–32.

8. Dakhel A. M. et al. Github copilot ai pair programmer: Asset or liability. Journal of Systems and Software, 2023, vol. 203, pp. 111–734.

9. Ye X. W. et al. Copyright Protection and Risk Mitigation for AI-Generated Content (AIGC) Creations. Journal of Computers, 2025, vol. 36, no. 5, pp. 167–181.

10. Shi Y. Study on security risks and legal regulations of generative artificial intelligence. Science of law journal, 2023, vol. 2, no. 11, pp. 17–23.

11. Atkinson D., Morrison J. A legal risk taxonomy for generative artificial intelligence. arXiv preprint arXiv:2404.09479, 2024, pp. 1–25.

12. George A. S. Preparing students for an AI-driven world: Rethinking curriculum and pedagogy in the age of artificial intelligence. Partners Universal Innovative Research Publication, 2023, vol. 1, no. 2, pp. 112–136.

13. Baidoo-Anu D., Ansah L. O. Education in the era of generative artificial intelligence (AI): Understanding the potential benefits of ChatGPT in promoting teaching and learning. Journal of AI, 2023, vol. 7, no. 1, pp. 52–62.

14. Hellas A. et al. Exploring the responses of large language models to beginner programmers’ help requests. Proceedings of the 2023 ACM Conference on International Computing Education Research. 2023. Vol. 1, pp. 93–105.

15. Uppalapati V. K., Nag D. S. A comparative analysis of AI models in complex medical decision-making scenarios: evaluating ChatGPT, Claude AI, Bard, and Perplexity. Cureus, 2024, vol. 16, no. 1.

16. Aydin O. et al. Generative ai in academic writing: A comparison of deepseek, qwen, chatgpt, gemini, llama, mistral, and gemma. arXiv preprint arXiv:2503.04765, 2025, pp. 1–21.

17. Singh A. et al. Openai gpt-5 system card. arXiv preprint arXiv:2601.03267, 2025, pp. 1–19.

18. Martins S. Artificial intelligence-assisted classification of library resources: The case of Claude AI. Artificial Intelligence, 2024, vol. 2, pp. 27.

19. Baninemeh E. et al. A security risk assessment method for Distributed Ledger Technology (DLT) based applications: three industry case studies. arXiv preprint arXiv:2401.12358, 2024, pp. 1–39.

20. Xu Z., Sheng V. S. Detecting AI-generated code assignments using perplexity of large language models. Proceedings of the aaai conference on artificial intelligence, 2024, vol. 38, no. 21, pp. 23155–23162.