Maksim V. Sergievsky, Ph. D., Docent, Assistance Professor of Department Cybernetics, National Research Nuclear University MEPhI, https://orcid.org/0009-0005-8528-9177, e-mail: sermax@yandex.ru
Aleksey I. Vinokur, Dr. Sc. (Technic), Professor, Department Informatica and Information, Technology Moscow State Polytechnic University, https://orcid.org/0000-0002-6914-2520, e-mail: alex.vinokour@gmail.com
It is safe to say, that whatever the attitude to technology artificial intelligence (AI), they will be in demand in the learning process in the near future. This article evaluates the potential of using AI systems, in particular neural networks of GPT-class, in teaching programming in algorithmic languages. The authors conducted a study in which three AI systems — ChatGPT-4, Gemini Ultra and GigaChat — were used to solve a number of model tasks in order to generate code in Python and Scala languages. The results of AI systems were evaluated according to criteria such as running time, code efficiency, and comment quality. The conducted research has shown that, despite the generally optimistic attitude towards the use of AI in programming teaching, it is necessary to maintain caution and a reasonable balance of using both classical teaching methods and artificial intelligence systems.
Keywords: programming, teaching, artificial intelligence, programming code generation, code quality
References
1. Elsen-Rooney, M. NYC education department blocks ChatGPT on school devices, networks. Chalkbeat. 2023, Jan. 4. URL: https://ny.chalkbeat.org/2023/1/3/23537987/nyc-schools-ban-chatgpt-writi...
2. Sazonov, A.P. Use of AI in programming. Universum: Technical Sciences. 2024. V. 3 (120) [Electronic resource]. DOI: 10.32743/UniTech.2024.120.3.17010, URL: https://7universum.com/ru/tech/archive/item/17010
3. Marcus, G., Davis, E. Artificial Intelligence: Rebooting. How to create a machine intelligence that can really be trusted (Biblioteka Sbera: Artificial Intelligence). Moscow: Publ. House Intellectual Literature, 2021. 328 p. URL: https://sberuniversity.ru/research/biblio/10432/ (accessed on: 18.06.2024).
4. Lapan, М. Deep Reinforcement Learning Hands-On. eBook. Published by Packt, 2018.
5. Sergievsky, G., Sergievsky, M. Conception and Linguistic Means of Representation and Knowledge Processing at the Semantic Level. Automatic Documentation and Mathematical Linguistics. 2023. V. 57. Nо. 2. P. 127–133.
6. Zaytsev K.S., Sergievsky, M.V. Using foreign experience in the preparation of master’s degree programs in IT. Alma Mater (Vestnik vysshey shkoly). 2014. No. 5. P. 62–67.
7. Farley, D. Modern Software Engineering: Doing What Works to Build Better Software Faster. Addison-Wesley, 2021. 256 p. ISBN 978-0137314911 Open Library OL34779880M
8. Eltarenko, E.A.; Sergievsky, M.V. Evaluation of hardware and pro-software by a multilevel system of criteria. Computer Press. 1998. No. 8. P. 268–272.
9. The future of AI-assisted programming — first examples. 2023. [Electronic resource]. URL: https://habr.com/ru/companies/timeweb/articles/745442/
- Writing Code with AI. [Электронный ресурс]. URL: https://docs.superblocks.com/generative-ai/writing-code-with-ai
.png)
