OCENA SYSTEMÓW AI OPARTYCH NA LLM NA EGZAMINIE MATURALNYM Z MATEMATYKI: IMPLIKACJE DLA ZARZĄDZANIA SZTUCZNĄ INTELIGENCJĄ I ZAPEWNIANIA JAKOŚCI W ORGANIZACJACH EDUKACYJNYCH

Autor

  • Maciej Renz Wrocław University of Economics and Business, Faculty of Management Autor
  • Grzegorz Borkowski Wrocław University of Economics and Business, Faculty of Management Autor
  • Ewelina Ząb Wrocław University of Economics and Business, Faculty of Management Autor

DOI:

https://doi.org/10.17512/znpcz.2026.3.10

Słowa kluczowe:

zarządzanie AI (AI governance), analiza błędów, systemy AI oparte na LLM, rozumowanie matematyczne, edukacja matematyczna, zapewnianie jakości

Abstrakt

Systemy AI oparte na dużych modelach językowych (LLM – Large Language Model) są coraz częściej wykorzystywane w edukacji, między innymi jako narzędzia do rozwiązywania, wyjaśniania i sprawdzania zadań matematycznych. W tym eksploracyjnym badaniu oceniono skuteczność czterech powszechnie używanych, dostępnych dla użytkowników systemów AI (ChatGPT, Claude, Gemini i Perplexity) na maturze z matematyki na poziomie podstawowym i rozszerzonym. Systemy testowano przez ich interfejsy czatu z użyciem oficjalnych arkuszy CKE, przekazywanych jako kompletne załączniki, oraz jednego, stałego promptu dla każdego poziomu. Odpowiedzi oceniano według oficjalnych zasad oceniania CKE, przyznając punkty za odpowiedzi końcowe oraz za poprawnie wykonane etapy zadań otwartych. Wszystkie systemy uzyskały wysokie wyniki na poziomie podstawowym (86-96%), natomiast na poziomie rozszerzonym wyniki były znacznie bardziej zróżnicowane (38-100%). Na poziomie podstawowym większość utraconych punktów wynikała z błędów w interpretacji wykresów, rysunków i innych informacji wizualnych. Na poziomie rozszerzonym błędy częściej wiązały się z rozumowaniem matematycznym lub doborem metody. Wyniki pokazują wartość uzupełniania łącznych wyników egzaminacyjnych o diagnostyczne profile błędów, rozróżniające przetwarzanie danych wizualnych od rozumowania matematycznego. Dla organizacji edukacyjnych takie benchmarki mogą stanowić jedno ze źródeł dowodów na potrzeby nadzoru ludzkiego, identyfikacji ryzyka i okresowej oceny jakości. 
Wyniki odzwierciedlają jednak działanie kompletnych systemów AI w określonych warunkach interfejsu i nie powinny być interpretowane jako bezpośredni dowód ich ogólnej skuteczności edukacyjnej. Potrzebne są dalsze badania obejmujące większą liczbę arkuszy egzaminacyjnych, powtarzanych prób oraz użytkowników edukacyjnych.

Bibliografia

Ahn, J., Verma, R., Lou, R., Liu, D., Zhang, R., & Yin, W. (2024). Large language models for mathematical reasoning: Progresses and challenges. Proceedings of the 18th Conference of the European Chapter of the Association for Computational Linguistics: Student Research Workshop, 225-237. https://doi.org/10.18653/v1/2024.eacl-srw.17

Bommasani, R., Hudson, D. A., Adeli, E., Altman, R., Arora, S., Arx, S. V., Bernstein, M. S., Bohg, J., Bosselut, A., Brunskill, E., Brynjolfsson, E., Buch, S., Card, D., Castellon, R., Chatterji, N. S., Chen, A. S., Creel, K. A., Davis, J., Demszky, D., … Liang, P. (2021). On the opportunities and risks of foundation models. arXiv. https://arxiv.org/abs/2108.07258

Brown, T., Mann, B., Ryder, N., Subbiah, M., Kaplan, J., Dhariwal, P., Neelakantan, A., Shyam, P., Sastry, G., Askell, A., Agarwal, S., Herbert-Voss, A., Krueger, G., Henighan, T., Child, R., Ramesh, A., Ziegler, D., Wu, J., Winter, C., … Amodei, D. (2020). Language models are few-shot learners. arXiv. https://doi.org/10.48550/arXiv.2005.14165

Bubeck, S., Chandrasekaran, V., Eldan, R., Gehrke, J., Horvitz, E., Kamar, E., Lee, P., Lee, Y. T.,

Li, Y., Lundberg, S., Nori, H., Palangi, H., Ribeiro, M., & Zhang, Y. (2023). Sparks of artifi-cial general intelligence: Early experiments with GPT-4. arXiv. https://doi.org/10.48550/arXiv.2303.12712

Cobbe, K., Kosaraju, V., Bavarian, M., Chen, M., Jun, H., Kaiser, L., Plappert, M., Tworek, J.,

Hilton, J., Nakano, R., Hesse, C., & Schulman, J. (2021). Training verifiers to solve math word problems. arXiv. https://arxiv.org/abs/2110.14168

Davis, F. D. (1989). Perceived usefulness, perceived ease of use, and user acceptance of infor-mation technology. MIS Quarterly, 13(3), 319-340. https://doi.org/10.2307/249008

Dong, Y., Jiang, X., Liu, H., Jin, Z., & Li, G. (2024). Generalization or memorization: Data con-tamination and trustworthy evaluation for large language models. arXiv. https://arxiv.org/abs/2402.15938

Dwivedi, Y. K., Hughes, L., Ismagilova, E., Aarts, G., Coombs, C., Crick, T., Duan, Y., Dwivedi, R., Edwards, J., Eirug, A., Galanos, V., Ilavarasan, P. V., Janssen, M., Jones, P., Kar, A. K.,

Kizgin, H., Kronemann, B., Lal, B., Lucini, B., … Williams, M. D. (2021). Artificial intelli-gence (AI): Multidisciplinary perspectives on emerging challenges, opportunities, and agen-da for research, practice and policy. International Journal of Information Management, 57, Article 101994. https://doi.org/10.1016/j.ijinfomgt.2019.08.002

Floridi, L., Cowls, J., Beltrametti, M., Chatila, R., Chazerand, P., Dignum, V., Luetge, C., Madelin, R., Pagallo, U., Rossi, F., Schafer, B., Valcke, P., & Vayena, E. (2018). AI4People—An eth-ical framework for a good AI society: Opportunities, risks, principles, and recommenda-tions. Minds and Machines, 28, 689-707. https://doi.org/10.1007/s11023-018-9482-5

Guo, D., Yang, D., Zhang, H., Song, J., Wang, P., Zhu, Q., Xu, R., Zhang, R., Ma, S., Bi, X., Zhang, X., Yu, X., Wu, Y., Wu, Z. F., Gou, Z., Shao, Z., Li, Z., Gao, Z., Liu, A., … Zhang. Z. (2025). DeepSeek-R1 incentivizes reasoning in LLMs through reinforcement learning. Nature, 645, 633-638. https://doi.org/10.1038/s41586-025-09422-z

Hendrycks, D., Burns, C., Basart, S., Zou, A., Mazeika, M., Song, D. X., & Steinhardt, J. (2020). Measuring massive multitask language understanding. arXiv. https://arxiv.org/abs/2009.03300

Hendrycks, D., Burns, C., Kadavath, S., Arora, A., Basart, S., Tang, E., Song, D. X., & Steinhardt, J. (2021). Measuring mathematical problem solving with the MATH dataset. arXiv. https://arxiv.org/abs/2103.03874

Heung, Y., & Chiu, T. K. F. (2025). How ChatGPT impacts student engagement from a systematic review and meta-analysis study. Computers and Education: Artificial Intelligence. https://doi.org/10.1016/j.caeai.2025.100361

Jassem, K., Ciesiólka, M., Graliński, F., Jabłoński, P., Pokrywka, J., Kubis, M., Jabłońska, M.,

& Staruch, R. (2025). LLMzSzŁ: A comprehensive LLM benchmark for Polish. arXiv. https://arxiv.org/abs/2501.02266

Kaplan, J., McCandlish, S., Henighan, T., Brown, T. B., Chess, B., Child, R., Gray, S., Radford, A., Wu, J., & Amodei, D. (2020). Scaling laws for neural language models. arXiv. https://arxiv.org/abs/2001.08361

Kasneci, E., Seßler, K., Küchemann, S., Bannert, M., Dementieva, D., Fischer, F., Gasser, U., Groh, G. L., Günnemann, S., Hüllermeier, E., Krusche, S., Kutyniok, G., Michaeli, T., Nerdel, C., Pfeffer, J., Poquet, O., Sailer, M., Schmidt, A., Seidel, T., … Kasneci, G. (2023). ChatGPT for good? On opportunities and challenges of large language models for educa-tion. Learning and Individual Differences, 103, Article 102274. https://doi.org/10.1016/j.lindif.2023.102274

Kojima, T., Gu, S. S., Reid, M., Matsuo, Y., & Iwasawa, Y. (2022). Large language models are zero-shot reasoners. arXiv. https://arxiv.org/abs/2205.11916

Kotter, J. P. (1996). Leading change. Harvard Business School Press.

Lewkowycz, A., Andreassen, A., Dohan, D., Dyer, E., Michalewski, H., Ramasesh, V., Slone, A., Anil, C., Schlag, I., Gutman-Solo, T., Wu, Y., Neyshabur, B., Gur-Ari, G., & Misra, V. (2022). Solving quantitative reasoning problems with language models. arXiv. https://doi.org/10.48550/arXiv.2206.14858

Mirzadeh, I., Alizadeh-Vahid, K., Shahrokhi, H., Tuzel, O., Bengio, S., & Farajtabar, M. (2024). GSM-Symbolic: Understanding the limitations of mathematical reasoning in large language models. arXiv. https://arxiv.org/abs/2410.05229

Nonaka, I., & Takeuchi, H. (1995). The knowledge-creating company: How Japanese companies

create the dynamics of innovation. Oxford University Press. https://doi.org/10.1093/oso/9780195092691.001.0001

OpenAI, Achiam, J., Adler, S., Agarwal, S., Ahmad, L., Akkaya, I., Aleman, F. L., Almeida, D.,

Altenschmidt, J., Altman, S., Anadkat, S., Avila, R., Babuschkin, I., Balaji, S., Balcom, V.,

Baltescu, P., Bao, H., Bavarian, M., Belgum, J., … Zoph, B. (2023). GPT-4 technical report. arXiv. https://arxiv.org/abs/2303.08774

OpenAI, Jaech, A., Kalai, A., Lerer, A., Richardson, A., El-Kishky, A., Low, A., Helyar, A., Madry, A., Beutel, A., Carney, A., Iftimie, A., Karpenko, A., Passos, A. T., Neitz, A., Pro-kofiev, A., Wei, A., Allison Tam, Bennett, A., … Li, Z. (2024). OpenAI o1 system card. arXiv. https://arxiv.org/abs/2412.16720

Parasuraman, A., Zeithaml, V., & Berry, L. L. (1988). SERVQUAL: A multiple-item scale for measuring consumer perceptions of service quality. Journal of Retailing, 64(1), 12-40.

Pardos, Z. A., & Bhandari, S. (2024). ChatGPT-generated help produces learning gains equivalent to human tutor-authored help on mathematics skills. PLOS One, 19, Article e0304013. https://doi.org/10.1371/journal.pone.0304013

Pokrywka, J., Kaczmarek, J., & Gorzelańczyk, E. (2024). GPT-4 passes most of the 297 written Polish board certification examinations. SSRN. https://doi.org/10.2139/ssrn.4861528

Pyeon, G., Heo, I., Jung, J., Hwang, T., Namgoong, H., Seo, H., Han, Y., Kim, E., Kang, H., & Jung, S. (2025). Evaluating large language models on the 2026 Korean CSAT mathematics exam: Measuring mathematical ability in a zero-data-leakage setting. arXiv. https://arxiv.org/abs/2511.18649

Rogers, E. M. (2003). Diffusion of innovations (5th ed.). Free Press.

Rosoł, M., Gąsior, J. S., Łaba, J., Korzeniewski, K., & Młyńczak, M. (2023). Evaluation of the performance of GPT-3.5 and GPT-4 on the Polish Medical Final Examination. Scientific Re-ports, 13, Article 20512. https://doi.org/10.1038/s41598-023-46995-z

Setälä, M., Sikström, P., Heilala, V., & Kärkkäinen, T. (2025). Assessment of evolving large lan-guage models in upper secondary mathematics. Proceedings of ICETC 2025, 210-214. https://doi.org/10.1109/ICETC66579.2025.11387533

Turmuzi, M., Azmi, S., & Kertiyani, N. M. (2026). ChatGPT in school mathematics education: A systematic review of opportunities, challenges, and pedagogical implications. Teaching and Teacher Education, 170, Article 105286.

Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A. N., Kaiser, L., & Po-losukhin, I. (2017). Attention is all you need. Advances in Neural Information Processing Systems. https://arxiv.org/abs/1706.03762

Wei, J., Tay, Y., Bommasani, R., Raffel, C., Zoph, B., Borgeaud, S., Yogatama, D., Bosma, M., Zhou, D., Metzler, D., Chi, E. H., Hashimoto, T., Vinyals, O., Liang, P., Dean, J., & Fedus, W. (2022b). Emergent abilities of large language models. arXiv. https://arxiv.org/abs/2206.07682

Wei, J., Wang, X., Schuurmans, D., Bosma, M., Chi, E. H., Xia, F., Le, Q., & Zhou, D. (2022a). Chain-of-thought prompting elicits reasoning in large language models. arXiv. https://arxiv.org/abs/2201.11903

Zhao, W., Zhou, K., Junyi, L., Tianyi, T., Wang, X., Hou, Y., Min, Y., Zhang, B., Zhang, J., Dong, Z., Du, Y., Yang, C., Chen, Y., Chen, Z., Jiang, J., Ren, R., Li, Y., Tang, X., Liu, Z., … Wen, J. R. (2026). A survey of large language models. Frontiers of Computer Science, 20, Article 2012627. https://doi.org/10.1007/s11704-026-60308-3

Opublikowane

2026-09-30

Numer

Dział

Artykuły