UDC 378.14

Tadzhibaev R.R.¹ (Kazan, Russian Federation) – divinerustam@gmail.com; Kosmodemyanskaya S.S.¹ (Kazan, Russian Federation) – svetlanakos@mail.ru; Yarullin I.F.¹ (Kazan, Russian Federation) – yarullin_ilnar@mail.ru

¹Kazan Federal University

Abstract. The modern educational system faces the need to optimize and objectify the processes of assessing students’ knowledge, especially in subject areas that require processing specific content, such as chemistry. The growing workload of teachers actualizes the search for effective automation tools. The rapid development of artificial intelligence (AI), neural networks and large language models opens up significant prospects for solving this problem. This article analyzes the potential of using neural networks for automated assessment of students’ chemistry tests. A general methodology for such assessment is proposed, including the key stage of digitizing handwritten answers. The results of an experiment on the use of large language models ChatGPT for digitizing real handwritten chemistry tests are presented, demonstrating high accuracy of recognition of specific chemical content (formulas, equations, calculations). The success of this stage removes one of the main technical barriers to full automation. The benefits of implementing AI systems (speed, scalability, potential objectivity) are discussed, as well as significant challenges and limitations (accuracy of complex case analysis, data requirements, interpretability, assessment of creative responses, bias). The critical role of accurately digitizing handwritten responses containing complex chemical notation is particularly emphasized. Special attention is paid to ethical aspects, including data privacy, transparency, fairness of assessment, academic integrity, and maintaining the role of the teacher. The conclusion summarizes the main findings and identifies key areas for further research to create comprehensive and ethically responsible automated chemistry assessment systems.

Keywords: chemistry, teaching methods, artificial intelligence, neural networks, large language models, ChatGPT, educational assessment.

References:

1. Zufarova A.S. Integration of Artificial Intelligence into Modern Education: Advantages and Challenges. CITISE, 2025, no. 1, pp. 652-659. (In Russian). URL: https://elibrary.ru/vmftaa
2. Subbotina M.V. Artificial Intelligence and Higher Education – Enemies or Allies. Bulletin of the Peoples’ Friendship University of Russia. Series: Sociology, 2024, vol. 24, no. 1, pp. 176-183. (In Russian). URL: https://www.elibrary.ru/xuuytg
3. Yamtinah S., Wiyarsi A., Widarti H.R. Fine-tuning AI models for enhanced consistency and precision in chemistry educational assessments. Computers and Education: Artificial Intelligence, 2025, vol. 8, 100399. DOI: 10.1016/j.caeai.2025.100399
4. Calatayud V.G., Espinosa M.P.P., Vila R.R. “Artificial intelligence for student assessment: A systematic review”. Applied Sciences, 2021, vol. 11(12), 5467. DOI: 10.3390/app11125467
5. Reddy M.R., Walter N.G., Sevryugina Y.V. Implementation and Evaluation of a ChatGPT-Assisted Special Topics Writing Assignment in Biochemistry. Journal of Chemical Education, 2024, vol. 101, issue 7, pp. 2740–2748. DOI: 10.1021/acs.jchemed.4c00226
6. Du Y., Duan C., Bran A.M. et al. Large Language Models are Catalyzing Chemistry Education. ChemRxiv, 2024. DOI: 10.26434/chemrxiv-2024-h722v
7. Hassija V., Chamola V., Mahapatra A. et al. “Interpreting Black-Box Models: A Review on Explainable Artificial Intelligence”. Cognitive Computation, 2024, vol. 16, pp. 45–74. DOI: 10.1007/s12559-023-10179-8
8. Zhang K., Chai C.S., Wang Z. et al. The application of AI technologies in STEM education: a systematic review from 2011 to 2021. STEM Education Review, 2022, vol. 59. DOI: 10.1186/s40594-022-00377-5
9. Ruff E.F., Zemke J.M.O. Discussing the Ethics of Professional AI Use in Undergraduate Chemistry Courses. Journal of Chemical Education, 2024, vol. 102, Issue 4. DOI: 10.1021/acs.jchemed.4c01038
10. Blonder R., Feldman-Maggor Y. AI for chemistry teaching: responsible AI and ethical considerations. Chemistry Teacher International, 2024, vol. 6, issue 4. DOI: 10.1515/cti-2024-0014
11. Yuriev E., Wink D.J., Holme T.A. The Dawn of Generative Artificial Intelligence in Chemistry Education. Journal of Chemical Education, 2024, vol. 101, issue 8, pp. 2957-2959. DOI: 10.1021/acs.jchemed.4c00836
12. Sosnin S., Garanin A., Kurakin M. Deep learning Transformer model for recognition of chemical formulas in scientific articles. Chemistry-Methods, 2022, vol. 3, issue 2. e202200042. (In Russian).
13. Tadjibaev R.R., Kosmodemyanskaya S.S. Analysis of the application of neural networks in chemical education. Proc. “Management of the development of functional literacy”. Kazan, Kazan University Publ., 2024. 366 p. (In Russian). URL: https://elibrary.ru/vctdmg
14. Tadjibaev R.R., Kosmodemyanskaya S.S. Inclusive education in the era of digitalization: the role of neural networks in the development of chemical education. Proc. “XIII International scientific and practical conference”. Kazan, Poznanie Publ., 2024. 428 p. (In Russian). URL: https://elibrary.ru/bqcymh
15. Gilmutdinova E.D., Nasibullov R.R., Yarullin I.F. Using Internet memes as a tool for the educational process: educational potential. Bulletin of pedagogical sciences, 2025, no. 1, pp. 58-65. (In Russian). DOI: 10.62257/2687-1661-2025-1-58-65

For citation: Tadjibaev R.R., Kosmodemyanskaya S.S., Yarullin I.F. Analysis of automated assessment using neural networks in chemistry using ChatGPT. CITISE, 2025, no. 2, pp. 818-828.