Informacionnye TehnologiiInformacionnye TehnologiiИнформационные технологии1684-6400New Technologies Publishing House70602010.17587/it.32.185-194Intelligent systems and technologiesИнтеллектуальные системы и технологииReview ArticleFrom the unknown to transparency: a review of information technologies in explainable AIОт неизвестности к прозрачности: обзор технологий объяснимого ИИ (XAI)AvdoshinS. M.АвдошинС. М.
Russian Federation

Cand. Tech. Sc., Professor

канд. техн. наук, проф.

savdoshin@hse.ruPesotskayaE. Y.ПесоцкаяЕ. Ю.
Russian Federation

Cand. Econ. Sc., Ass. Professor

канд. экон. наук, доц.

epesotskaya@hse.ruHSE University — National Research University Higher School of EconomicsНациональный исследовательский университет "Высшая школа экономики"110420263241851941104202611042026Copyright ©; 2026, Informacionnye TehnologiiCopyright ©; 2026, Информационные технологии2026Informacionnye TehnologiiИнформационные технологииhttps://journals.eco-vector.com/1684-6400/article/view/706020

With the rapid advancement of artificial intelligence, and deep learning in particular, models have emerged that are capable of delivering highly accurate predictions. However, the internal logic of such models remains difficult to interpret—an issue of critical importance, especially in domains where the correctness of an algorithm directly affects high-stakes decision-making. One promising avenue for addressing this challenge is Explainable Artificial Intelligence (XAI), which focuses on developing approaches that clarify model behavior and provide transparent reasoning behind the results obtained. This work examines theoretical foundations of XAI, with particular attention to the classification of methods and the challenges posed by the "black box" nature of machine learning models. The review highlights the necessity of advancing new XAI techniques, outlines potential ways to reconcile high predictive accuracy with sufficient interpretability, and lays the groundwork for further research in this field.

С развитием ИИ, и в особенности глубокого обучения, появились модели, способные давать крайне точные прогнозы. Однако их внутренняя логика остается трудной для понимания — и это серьезная проблема, особенно в сферах, где от корректности алгоритма зависят критически важные решения. Одним из перспективных путей ее решения считается направление Explainable Artificial Intelligence (XAI) — разработка подходов, позволяющих прояснять работу моделей и объяснять логику полученных результатов.

В рамках работы обсуждаются теоретические подходы к XAI, отдельное внимание уделяется классификации методов, трудностям, связанным с "черным ящиком" в машинном обучении. Проведенный обзор подчеркивает необходимость развития новых моделей XAI, описывает, каким образом можно совместить высокую точность моделей с достаточной степенью интерпретируемости, и формирует основания для дальнейших исследований в данной области.

explainable artificial intelligenceXAImachine learninginterpretabilitytrusttransparencyобъяснимый искусственный интеллектXAIмашинное обучениеинтерпретируемостьдовериепрозрачностьSamek W., Wiegand T., Müller K.-R. Explainable artificial intelligence: Understanding, visualizing and interpreting deep learning models, ITU Journal: ICT Discoveries, 2018, vol. 1, no. 1, pp. 39—48.Samek W., Wiegand T., Müller K.-R. Explainable artificial intelligence: Understanding, visualizing and interpreting deep learning models // ITU Journal: ICT Discoveries. 2018. Vol. 1, N. 1. P. 39—48.Barredo Arrieta A., Díaz-Rodríguez N., Del Ser J. et al. Explainable artificial intelligence (XAI): Concepts, taxonomies, opportunities and challenges toward responsible AI, Information Fusion, 2020, vol. 58, pp. 82—115.Barredo Arrieta A., Díaz-Rodríguez N., Del Ser J. et al. Explainable artificial intelligence (XAI): Concepts, taxonomies, opportunities and challenges toward responsible AI // Information Fusion. 2020. Vol. 58. P. 82—115.Guidotti R., Monreale A., Ruggieri S. et al. A survey of methods for explaining black box models, ACM Computing Surveys, 2018, vol. 51, no. 5, art. 93.Guidotti R., Monreale A., Ruggieri S. et al. A survey of methods for explaining black box models // ACM Computing Surveys. 2018. Vol. 51, N. 5. Art. 93.Gunning D., Stefik M., Choi J. et al. DARPA’s explainable AI (XAI) program: A retrospective, Applied AI Letters, 2021, vol. 2, no. 4. e61.Gunning D., Stefik M., Choi J. et al. DARPA’s explainable AI (XAI) program: A retrospective // Applied AI Letters. 2021. Vol. 2, N. 4. e61.Adadi A., Berrada M. Peeking inside the black-box: A survey on explainable artificial intelligence (XAI), IEEE Access, 2018, vol. 6, pp. 52138—52160.Adadi A., Berrada M. Peeking inside the black-box: A survey on explainable artificial intelligence (XAI) // IEEE Access. 2018. Vol. 6. P. 52138—52160.Murdoch W. J., Singh C., Kumbier K. et al. Definitions, methods, and applications in interpretable machine learning, Proceedings of the National Academy of Sciences, 2019, vol. 116, no. 44, pp. 22071—22080.Murdoch W. J., Singh C., Kumbier K. et al. Definitions, methods, and applications in interpretable machine learning // Proceedings of the National Academy of Sciences. 2019. Vol. 116, N. 44. P. 22071—22080.Doshi-Velez F., Kim B. Towards a rigorous science of interpretable machine learning, arXiv:1702.08608, 2017.Doshi-Velez F., Kim B. Towards a rigorous science of interpretable machine learning // arXiv:1702.08608. 2017.Tjoa E., Guan C. A survey on explainable artificial intelligence (XAI): Toward medical XAI, IEEE Transactions on Neural Networks and Learning Systems. 2021, vol. 32, no. 11, pp. 4793—4813.Tjoa E., Guan C. A survey on explainable artificial intelligence (XAI): Toward medical XAI // IEEE Transactions on Neural Networks and Learning Systems. 2021. Vol. 32, N. 11. P. 4793—4813.Lipton Z. C. The mythos of model interpretability, arXiv: 1606.03490, 2016.Lipton Z. C. The mythos of model interpretability // arXiv:1606.03490. 2016.Tabassi E. (ed.). Artificial intelligence risk management framework (AI RMF 1.0), National Institute of Standards and Technology (NIST), 2023.Tabassi E. (ed.) Artificial intelligence risk management framework (AI RMF 1.0). National Institute of Standards and Technology (NIST) AI 100-1. 2023.Mylrea M., Robinson N. Artificial intelligence (AI) trust framework and maturity model, Entropy, 2023, vol. 25, no. 10, art. 1429.Mylrea M., Robinson N. Artificial intelligence (AI) trust framework and maturity model // Entropy. 2023. Vol. 25, Т. 10. Art. 1429.Molnar C. Interpretable machine learning: A guide for making black box models explainable, 2022, 317 p., Self-published, free PDF version, available at: https://originalstatic.aminer.cn/misc/pdf/Molnar-interpretable-machine-learning_compressed.pdf (accessed: 5 Feb 2026).Molnar C. Interpretable machine learning: A guide for making black box models explainable. 2022. 317 p. [Электронный ресурс] URL: https://originalstatic.aminer.cn/misc/pdf/Molnar-interpretable-machine-learning_compressed.pdf (дата обращения: 5 февраля 2026).Ghosh D., Dhingra D., Hassija V. et al. A review on trustworthy and explainable artificial intelligence (XAI), IEEE Access, 2023, vol. 11, pp. 78994—79015.Ghosh D., Dhingra D., Hassija V. et al. A review on trustworthy and explainable artificial intelligence (XAI) // IEEE Access. 2023. Vol. 11. P. 78994—79015.Phillips P. J., Hahn C. A., Fontana P. C. et al. Four principles of explainable artificial intelligence, National Institute of Standards and Technology, Internal Report 8312, 2021, 37 p., DOI: 10.6028/NIST.IR.8312.Phillips P. J., Hahn C. A., Fontana P. C. et al. Four principles of explainable artificial intelligence. National Institute of Standards and Technology, Internal Report 8312. 2021. 37 p. DOI: 10.6028/NIST.IR.8312Gregor K., LeCun Y. Learning fast approximations of sparse coding, Proceedings of ICML, 2010, pp. 1—8.Gregor K., LeCun Y. Learning fast approximations of sparse coding // Proceedings of ICML. 2010. P. 1—8.Springenberg J. T., Dosovitskiy A., Brox T., Riedmiller M. Striving for simplicity: The all convolutional net, arXiv:1412.6806, 2015.Springenberg J. T., Dosovitskiy A., Brox T., Riedmiller M. Striving for simplicity: The all convolutional net // arXiv:1412.6806. 2015.Smilkov D., Thorat N., Kim B. et al. SmoothGrad: Removing noise by adding noise, arXiv:1706.03825, 2017.Smilkov D., Thorat N., Kim B. et al. SmoothGrad: Removing noise by adding noise // arXiv:1706.03825. 2017.Zhou B., Khosla A., Lapedriza A. et al. Learning deep features for discriminative localization, Proceedings of CVPR, 2016, pp. 2921—2929.Zhou B., Khosla A., Lapedriza A. et al. Learning deep features for discriminative localization // Proceedings of CVPR. 2016. P. 2921—2929.Selvaraju R. R., Cogswell M., Das A. et al. Grad-CAM: Visual explanations from deep networks via gradient-based localization, Proceedings of ICCV, 2017, pp. 618—626.Selvaraju R. R., Cogswell M., Das A. et al. Grad-CAM: Visual explanations from deep networks via gradient-based localization // Proceedings of ICCV. 2017. P. 618—626.Ribeiro M. T., Singh S., Guestrin C. ‘Why should I trust you?’: Explaining the predictions of any classifier, Proceedings of KDD, 2016, pp. 1135—1144.Ribeiro M. T., Singh S., Guestrin C. ‘Why should I trust you?’: Explaining the predictions of any classifier // Proceedings of KDD. 2016. P. 1135—1144.Guo W., Mu D., Zhang X. et al. LEMNA: Explaining deep learning based security applications, Proceedings of CCS, 2018, pp. 364—379.Guo W., Mu D., Zhang X. et al. LEMNA: Explaining deep learning based security applications // Proceedings of CCS. 2018. P. 364—379.Lundberg S. M., Lee S.-I. A unified approach to interpreting model predictions, Advances in Neural Information Processing Systems, 2017, vol. 30, pp. 4765—4774.Lundberg S. M., Lee S.-I. A unified approach to interpreting model predictions // Advances in Neural Information Processing Systems. 2017. Vol. 30. P. 4765—4774.Ozturkkal B., Wahlström R. R. Explaining mortgage defaults using SHAP and LASSO, Computational Economics, 2024 (online first).Ozturkkal B., Wahlström R. R. Explaining mortgage defaults using SHAP and LASSO // Computational Economics. 2024. (online first).Kapranov V. V., Tugaenko V. Yu. A machine learning approach for transmitted power estimation in power beaming systems, Informacionnye Tehnologii, 2023, vol. 29, no. 3, pp. 126—135.Kapranov V. V., Tugaenko V. Yu. A machine learning approach for transmitted power estimation in power beaming systems // Информационные технологии. 2023. Т. 29, № 3. С. 126—135.Mishra P. Practical Explainable AI Using Python, Moscow, DMK Press, 2022, 298 p.Мишра П. Объяснимые модели искусственного интеллекта на Python: Практическое руководство / Пер. с англ. С. В. Минц. М.: ДМК Пресс, 2022. 298 с.Abdul A., Vermeulen J., Wang D. et al. Trends and trajectories for explainable, accountable and intelligible systems, Proceedings of CHI, 2018, Paper 582, 18 p.Abdul A., Vermeulen J., Wang D. et al. Trends and trajectories for explainable, accountable and intelligible systems // Proceedings of CHI. 2018. Paper 582. 18 p.European Data Protection Supervisor. EDPS TechDispatch 2/2023 — Explainable artificial intelligence, Luxembourg, Publications Office of the European Union, 2023.European Data Protection Supervisor. EDPS TechDispatch 2/2023 — Explainable artificial intelligence. Luxembourg: Publications Office of the European Union, 2023.Hamon R., Junklewitz H., Sanchez I. Robustness and explainability of artificial intelligence: From technical to policy solutions, Luxembourg,: Publications Office of the European Union, 2020.Hamon R., Junklewitz H., Sanchez I. Robustness and explainability of artificial intelligence: From technical to policy solutions. Luxembourg: Publications Office of the European Union, 2020.Avdoshin S. M., Pesotskaya E. Yu., Patrushev K. A. Trusted artificial intelligence technologies, Informacionnye Tehnologii, 2024, vol. 30, no. 8, pp. 400—410.Авдошин С. М., Песоцкая Е. Ю., Патрушев К. А. Технологии доверенного искусственного интеллекта // Информационные технологии. 2024. Т. 30, № 8. С. 400—410.