voprstatВопросы статистикиVoprosy Statistiki2313-63832658-549910.34023/2313-6383-2025-32-2-5-14voprstat-1885Research ArticleОРГАНИЗАЦИЯ И РАЗВИТИЕ ГОСУДАРСТВЕННОЙ СТАТИСТИКИORGANIZATION AND DEVELOPMENT OF STATE STATISTICSТехнологии искусственного интеллекта в официальной статистике: возможности использования и рискиArtificial Intelligence Technologies in Official Statistics: Use Cases and Riskshttps://orcid.org/0000-0003-1253-4029БашинаО. Э.BashinaO. E.

Башина Ольга Эмильевна – д-р экон. наук, профессор, профессор кафедры прикладной информатики и статистики

111395, г. Москва, ул. Юности, д. 5

Olga E. Bashina – Dr. Sci. (Econ.), Professor; Professor, Department of Applied Computer Science and Statistics

5, Yunosti Str., Moscow, 111395

bashina_o_e@mail.ruhttps://orcid.org/0000-0002-9080-7953МатраеваЛ. В.MatraevaL. V.

Матраева Лилия Валериевна – д-р экон. наук, профессор, профессор кафедры финансового учета и контроля Института кибербезопасности и цифровых технологий

107996, г. Москва, ул. Стромынка, д. 20

Liliia V. Matraeva – Dr. Sci. (Econ.), Professor; Professor, Department of Financial Accounting and Control, Institute for Cybersecurity and Digital Technologies

20, Stromynka Str., Moscow, 107996

lilia.matraeva@gmail.comhttps://orcid.org/0000-0001-8707-1642ВасютинаЕ. С.VasiutinaE. S.

Васютина Екатерина Сергеевна – канд. экон. наук, доцент, доцент кафедры финансового учета и контроля Института кибербезопасности и цифровых технологий

107996, г. Москва, ул. Стромынка, д. 20

Ekaterina S. Vasiutina – Cand. Sci. (Econ.), Associate Professor; Associate Professor, Department of Financial Accounting and Control, Institute for Cybersecurity and Digital Technologies

20, Stromynka Str., Moscow, 107996

esvas@mail.ruМосковский гуманитарный университетРоссияMoscow University for the HumanitiesRussian FederationМИРЭА – Российский технологический университетРоссияMIREA – Russian Technological University (RTU MIREA)Russian Federation202530042025322514Copyright © Башина О.Э., Матраева Л.В., Васютина Е.С., 20252025Башина О.Э., Матраева Л.В., Васютина Е.С.Bashina O.E., Matraeva L.V., Vasiutina E.S.Данная работа распространяется под лицензией Creative Commons Attribution 4.0.This work is licensed under a Creative Commons Attribution 4.0 License.https://voprstat.elpub.ru/jour/article/view/1885

Статья посвящена исследованию возможностей и рисков использования генеративного искусственного интеллекта (Ген ИИ) в деятельности статистических служб. В условиях цифровизации и увеличения объемов данных Ген ИИ становится ключевым инструментом для автоматизации процессов сбора и обработки информации, оптимизации аналитики, а также повышения точности прогнозов. Однако внедрение этих технологий сопряжено с рядом значительных рисков, таких как утечка данных, киберугрозы, снижение доверия к официальной статистической информации и отсутствие единых стандартов оценки качества данных.Цель исследования – на основе анализа использования в деятельности статистических служб технологий искусственного интеллекта выявить потенциальные риски применения Ген ИИ и предложить рекомендации для их минимизации. Результаты проведенного исследования показывают, что главные угрозы связаны с утечкой данных, возможным получением недостоверной информации, а также с кибератаками на модели Ген ИИ. В статье предлагаются стратегии управления рисками, включая разработку политики использования ИИ, повышение прозрачности алгоритмов и создание систем мониторинга безопасности.Новизна работы заключается в комплексном анализе рисков Ген ИИ в деятельности статистических служб, что не получило ранее достаточного освещения в научной литературе. В отличие от предыдущих публикаций акцент сделан на институциональных аспектах и вопросах кибербезопасности, а также на необходимости разработки международных стандартов валидации данных и управления рисками

This article explores the risks and opportunities of using generative artificial intelligence (GenAI) in the activities of statistical agencies. In the context of digitalization and the growing volume of data, GenAI is becoming a key tool for automating data collection and processing, optimizing analytics and increasing the accuracy of forecasts. However, implementing these technologies poses several significant risks, such as data leakage, cybersecurity threats, reduced trust in official statistics, and the lack of unified standards for data quality assessment.The purpose of the study is to identify potential risks of using GenAI and to propose recommendations for minimizing them by analyzing the use of artificial intelligence technologies in the work of statistical agencies. The results show that the most significant threats are related to data leakage, possibility of receiving inaccurate information, and cyberattacks on GenAI models. The article suggests risk management strategies, including developing AI usage policies, increasing algorithm transparency, and creating security monitoring systems.The novelty of this work lies in the comprehensive analysis of GenAI risks in the activities of statistical agencies, which has not been sufficiently covered in the scientific literature before. Unlike previous publications, the emphasis is placed on institutional and cybersecurity aspects, as well as on the need to develop international standards for data validation and risk management

искусственный интеллектгенеративный искусственный интеллекткибербезопасностьрискиинновацииофициальная статистикавалидация данныхartificial intelligencegenerative artificial intelligence (GenAI)cybersecurityrisksinnovationsofficial statisticsdata validationReferencesChui M. et al. The State of AI in 2023: Generative AI’s Breakout Year. McKinsey & Company, 2023. URL: https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai-in-2023-generative-ais-breakout-year#/ (дата обращения 03.02.2025).Chui M. et al. The State of AI in 2023: Generative AI’s Breakout Year. McKinsey & Company; 2023. Available from: https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai-in-2023-generative-ais-breakout-year#/ (accessed 03.02.2025).Velkoff V. Modernizing the U.S. Census Bureau’s Demographic Programs // International Journal of Population Data Science. 2024. Vol. 9. No. 5. doi: https://doi.org/10.23889/ijpds.v9i5.2913.Velkoff V. Modernizing the U.S. Census Bureau’s Demographic Programs. International Journal of Population Data Science. 2024;9(5). Available from: https://doi.org/10.23889/ijpds.v9i5.2913.Lepage-Richer T., McKelvey F. States of Computing: On Government Organization and Artificial Intelligence in Canada // Big Data & Society. 2022. Vol. 9. Iss. 2. doi: https://doi.org/10.1177/20539517221123304.Lepage-Richer T., McKelvey F. States of Computing: On Government Organization and Artificial Intelligence in Canada. Big Data & Society. 2022;9(2). Available from: https://doi.org/10.1177/20539517221123304.Yang Y. et al. Machine Learning for Economic Forecasting: An Application to China's GDP Growth // Computation and Language. ArXiv. 2024. doi: https://doi.org/10.48550/arXiv.2407.03595.Yang Y. et al. Machine Learning for Economic Forecasting: An Application to China's GDP Growth. Computation and Language. ArXiv. 2024. Available from: https://doi.org/10.48550/arXiv.2407.03595.UNECE. Large Language Models for Official Statistics. HLG-MOS White Paper. December 2023. URL: https://unece.org/sites/default/files/2023-12/HLGMOS%20LLM%20Paper_Preprint_1.pdf (дата обращения 12.02.2025).UNECE. Large Language Models for Official Statistics. HLG-MOS White Paper. December 2023. Available from: https://unece.org/sites/default/files/2023-12/HLGMOS%20LLM%20Paper_Preprint_1.pdf (accessed 12.02.2025).UNECE. Use of Generative AI in Statistical Organizations: CES Survey Results. URL: https://unece.org/sites/default/files/2024-08/AI%20Survey%20results.pdf (дата обращения 12.02.2025).UNECE. Use of Generative AI in Statistical Organizations: CES Survey Results. Available from: https://unece.org/sites/default/files/2024-08/AI%20Survey%20results.pdf (accessed 12.02.2025).Матраева Л.В., Васютина Е.С., Башина О.Э. Методологические проблемы развития информационно-аналитической инфраструктуры для оценки состояния и прогнозирования сферы искусственного интеллекта // Проблемы прогнозирования. 2024. № 1. С. 116–131. doi: https://doi.org/10.47711/0868-6351-202-116-131.Matraeva L.V., Vasyutina E.S., Bashina O.E. Methodological Problems of Information Development-Analytical Infrastructure for Assessing the State and Forecasting the Sphere of Artificial Intelligence. Studies on Russian Economic Development. 2024;(1):116–131. Available from: https://doi.org/10.47711/0868-6351-202-116-131.Elsner M., Atkinson G., Zahidi S. The Global Risks Report 2025. 20th edition. Cologny/Geneva: World Economic Forum, 2024. URL: https://www.weforum.org/publications/global-risks-report-2025/ (дата обращения 12.02.2025).Elsner M., Atkinson G., Zahidi S. The Global Risks Report 2025. 20 th edition. Cologny, Geneva, Switzerland: World Economic Forum; 2024. Available from: https://www.weforum.org/publications/global-risks-report-2025/ (accessed 12.02.2025).Shabsigh G., Boukherouaa E.B. Generative Artificial Intelligence in Finance // Fintech Notes. NOTE/2023/006. Washington, DC: IMF, 2023. URL: https://www.imf.org/en/Publications/fintech-notes/Issues/2023/08/18/Generative-Artificial-Intelligence-in-Finance-Risk-Considerations-537570.Shabsigh G., Boukherouaa E.B. Generative Artificial Intelligence in Finance. Fintech Notes. NOTE/2023/006Washington, DC: IMF; 2023. Available from: https://www.imf.org/en/Publications/fintech-notes/Issues/2023/08/18/Generative-Artificial-Intelligence-in-Finance-Risk-Considerations-537570.Stanford University Human-Centered Artificial Intelligence. Artificial Intelligence Index Report 2024. URL: https://aiindex.stanford.edu/wp-content/uploads/2023/04/HAI_AI-Index-Report_2023.pdf (дата обращения 03.02.2025).Stanford University Human-Centered Artificial Intelligence. Artificial Intelligence Index Report 2024. Available from: https://aiindex.stanford.edu/wp-content/uploads/2023/04/HAI_AI-Index-Report_2023.pdf (accessed 03.02.2025).Zou A. et al. Universal and Transferable Adversarial Attacks on Aligned Language Models // Computation and Language. 2023. doi: doi.org/10.48550/arXiv.2307.15043.Zou A. et al. Universal and Transferable Adversarial Attacks on Aligned Language Models. Computation and Language. 2023. Available from: doi.org/10.48550/arXiv.2307.15043.Ozili P.K. Artificial Intelligence in Central Banking: Benefits and Risks of AI for Central Banks // Industrial Applications of Big Data, AI, and Blockchain. Hershey, PA: IGI Global Scientific Publishing, 2024. P. 70–82. doi: https://doi.org/10.4018/979-8-3693-1046-5.ch004.Ozili P.K. Artificial Intelligence in Central Banking: Benefits and Risks of AI for Central Banks. In: El Samad M. et al. (eds.) Industrial Applications of Big Data, AI, and Blockchain. Hershey, PA: IGI Global Scientific Publishing; 2024. P. 70–82. Available from: https://doi.org/10.4018/979-8-3693-1046-5.ch004.Perry N. et al. Do Users Write More Insecure Code with AI Assistants? // Proc. of the 2023 ACM SIGSAC Conf. on Computer and Communications Security (CCS '23). Copenhagen Denmark, November 26–30, 2023. New York, USA: Association for Computing Machinery, 2023. P. 2785– 2799. doi: https://doi.org/10.1145/3576915.3623157.Perry N. et al. Do Users Write More Insecure Code with AI Assistants? In: Proc. of the 2023 ACM SIGSAC Conference on Computer and Communications Security (CCS '23), Copenhagen Denmark, 26–30 November 2023. New York, USA: Association for Computing Machinery; 2023. P. 2785–2799. Available from: https://doi.org/10.1145/3576915.3623157.Levendowski A. How Copyright Law Can Fix AI’s Implicit Bias Problem // Washington Law Review. 2018. Vol. 93. No. 2. P. 579–630. URL: https://digitalcommons.law.uw.edu/wlr/vol93/iss2/2.Levendowski A. How Copyright Law Can Fix AI’s Implicit Bias Problem. Washington Law Review. 2018;93(2):579–630. Available from: https://digitalcommons.law.uw.edu/wlr/vol93/iss2/2.Кочергин С.В. и др. Кибербезопасность смарт-сетей: сравнение подходов машинного обучения для обнаружения аномалий // Russian Technological Journal. 2024. Т. 12. № 6. С. 7–19. doi: https://doi.org/10.32362/2500-316X-2024-12-6-7-19.Kochergin S.V. et al. Cybersecurity of Smart Grids: Comparison of Machine Learning Approaches Training for Anomaly Detection. Russian Technological Journal. 2024;12(6):7–19. (In Russ.) Available from: https://doi.org/10.32362/2500-316X-2024-12-6-7-19.Scherer M.U. Regulating Artificial Intelligence Systems: Risks, Challenges, Competencies, and Strategies // Harvard Journal of Law & Technology. 2016. Vol. 29. No. 2. doi: https://dx.doi.org/10.2139/ssrn.2609777.Scherer M.U. Regulating Artificial Intelligence Systems: Risks, Challenges, Competencies, and Strategies. Harvard Journal of Law & Technology. 2016;29(2). Available from: https://dx.doi.org/10.2139/ssrn.2609777.Mannes A. Governance, Risk, and Artificial Intelligence // AI Magazine. 2020. Vol. 41. Iss. 1. P. 61–69. doi: https://doi.org/10.1609/aimag.v41i1.5200Mannes A. Governance, Risk, and Artificial Intelligence. AI Magazine. 2020;41(1):61–69. Available from: https://doi.org/10.1609/aimag.v41i1.5200

The authors declare that there are no conflicts of interest present.