kulawrKutafin Law ReviewKutafin Law Review2713-05252713-0533MSAL10.17803/27130533.2026.2.36.269-295kulawr-883Research ArticleARTIFICIAL INTELLIGENCE, DIGITAL TECHNOLOGIES & DATA GOVERNANCEGoverning AI Foundation Models: Regulatory Approaches and Global Perspectiveshttps://orcid.org/0000-0003-3777-8927IzhaevO. A.

Osman A. Izhaev, Cand. Sci. (Law), Senior Research Fellow, School of Law and Administration

Tyumen

izhaev.osman@gmail.comhttps://orcid.org/0000-0003-1448-3085KuteynikovD. L.

Dmitry L. Kuteynikov, Cand. Sci. (Law), Senior Research Fellow, School of Law and Administration

Tyumen

kuteynikov@me.comUniversity of TyumenRussian Federation202617072026132269295Copyright © Izhaev O.A., Kuteynikov D.L., 20262026Izhaev O.A., Kuteynikov D.L.Izhaev O.A., Kuteynikov D.L.Данная работа распространяется под лицензией Creative Commons Attribution 4.0.This work is licensed under a Creative Commons Attribution 4.0 License.https://kulawr.msal.ru/jour/article/view/883

This article analyses the approaches of the European Union, the United States, and the Russian Federation to the legal regulation of the development and deployment of AI foundation models. Drawing on regulatory approaches reflected in the EU Artificial Intelligence Act, U.S. Executive Order 14110, the NIST AI Risk Management Framework, and relevant legal acts of the Russian Federation, it shows that legal regulation is primarily based on the classification of models according to their capabilities and associated risks.

Based on the findings derived from an analysis of the legal and ethical regulations in the Russian Federation, the study concludes that there are no specific rules dedicated to the development and deployment of foundation models.

To address this regulatory gap, the article formulates recommendations for policymakers in the Russian Federation. In particular, it proposes that foundation models be classified into two groups based on the computational resources required for their training. The first group includes models that do not pose significant risks, while the second group includes the most advanced models with potentially hazardous capabilities.

The proposed approach may contribute to the safe and secure development and deployment of foundation models in the Russian Federation by enabling a differentiated regulatory framework that takes into account their potential impact on human rights.

artificial intelligenceAI systemsfoundation modelsgeneral-purpose AI modelsfrontier modelslarge language modelstraining computeAI governancetransparencyEU AI ActThe study was supported by a grant from the Russian Science Foundation No. 25-28-01582. Available at: https://rscf. ru/project/25-28-01582/.ReferencesAlonocho, S. et al., (2025). Comparing Apples to Oranges: A Taxonomy for Navigating the Global Landscape of AI Regulation. In: Proceedings of June 23–26, 2025 (FAccT ‘25), USA, doi: 10.1145/3715275.3732059.Alonocho, S. et al., (2025). Comparing Apples to Oranges: A Taxonomy for Navigating the Global Landscape of AI Regulation. In: Proceedings of June 23–26, 2025 (FAccT ‘25), USA, doi: 10.1145/3715275.3732059.Bakhteev, D.V., (2024). The Legal Status of Artificial Intelligence Systems and a Model for Diɣerentiating Legal Liability for the Damage They Cause. Lex Russica, 4, pp. 9–23, doi: 10.17803/1729-5920.2024.209.4.009-023. (In Russ.).Bakhteev, D.V., (2024). The Legal Status of Artificial Intelligence Systems and a Model for Diɣerentiating Legal Liability for the Damage They Cause. Lex Russica, 4, pp. 9–23, doi: 10.17803/1729-5920.2024.209.4.009-023. (In Russ.).Balashova, A. and Lisycina, M., (2025). Russia will discuss banning AI with an alarming level of risk. Available at: https://www.rbc.ru/technology_and_media/11/04/2025/67f7dc399a79477fdd97bf30?from=from_main_3 [Accessed 15.08.2025]. (In Russ.).Balashova, A. and Lisycina, M., (2025). Russia will discuss banning AI with an alarming level of risk. Available at: https://www.rbc.ru/technology_and_media/11/04/2025/67f7dc399a79477fdd97bf30?from=from_main_3 [Accessed 15.08.2025]. (In Russ.).Bommasani, R. et al., (2021). On the Opportunities and Risks of Foundation Models. Available at: https://arxiv.org/abs/2108.07258v3 [Accessed 15.08.2025].Bommasani, R. et al., (2021). On the Opportunities and Risks of Foundation Models. Available at: https://arxiv.org/abs/2108.07258v3 [Accessed 15.08.2025].Brown, I., (2023). Allocating accountability in AI supply chains: a UKcentered regulatory perspective. Available at: https://www.adalovelaceinstitute.org/resource/ai-supply-chains/ [Accessed 15.08.2025].Brown, I., (2023). Allocating accountability in AI supply chains: a UKcentered regulatory perspective. Available at: https://www.adalovelaceinstitute.org/resource/ai-supply-chains/ [Accessed 15.08.2025].Bullock, C., van Arsdale, S. et al., (2024). Legal Considerations for Defining “Frontier Model”. Institute for Law & AI Working Paper, 2, pp. 1–28, doi: 10.2139/ssrn.4973370.Bullock, C., van Arsdale, S. et al., (2024). Legal Considerations for Defining “Frontier Model”. Institute for Law & AI Working Paper, 2, pp. 1–28, doi: 10.2139/ssrn.4973370.Caputo, N., (2025). Governing AI Beyond the Pretraining Frontier. Available at: https://arxiv.org/abs/2502.15719 [Accessed 15.08.2025].Caputo, N., (2025). Governing AI Beyond the Pretraining Frontier. Available at: https://arxiv.org/abs/2502.15719 [Accessed 15.08.2025].Casonato, C. and Olivato, G., (2024). AI Regulation in Europe: Exploring the Artificial Intelligence Act. In: Fabris, A. and Belardinelli, S. (eds). Digital Environments and Human Relations. Springer, pp. 87– 112, doi: 10.1007/978-3-031-76961-0_5.Casonato, C. and Olivato, G., (2024). AI Regulation in Europe: Exploring the Artificial Intelligence Act. In: Fabris, A. and Belardinelli, S. (eds). Digital Environments and Human Relations. Springer, pp. 87– 112, doi: 10.1007/978-3-031-76961-0_5.Cheng, H. et al., (2023). A Survey on Deep Neural Network PruningTaxonomy, Comparison, Analysis, and Recommendations. Available at: https://arxiv.org/abs/2308.06767 [Accessed 15.08.2025].Cheng, H. et al., (2023). A Survey on Deep Neural Network PruningTaxonomy, Comparison, Analysis, and Recommendations. Available at: https://arxiv.org/abs/2308.06767 [Accessed 15.08.2025].Chun, J. et al., (2024). Comparative Global AI Regulation: Policy Perspectives from the EU, China, and the U.S.. Available at: https://arxiv.org/abs/2410.21279 [Accessed 15.08.2025].Chun, J. et al., (2024). Comparative Global AI Regulation: Policy Perspectives from the EU, China, and the U.S.. Available at: https://arxiv.org/abs/2410.21279 [Accessed 15.08.2025].Datta, A., (2025). The Regulation of GPAI Model Providers Under the EU AI Act. In: Denga, M. and Hornuf, L. (eds). Regulatory Competition in the Digital Economy, pp. 119–138, doi: 10.1007/978-3-031-81089-3_6.Datta, A., (2025). The Regulation of GPAI Model Providers Under the EU AI Act. In: Denga, M. and Hornuf, L. (eds). Regulatory Competition in the Digital Economy, pp. 119–138, doi: 10.1007/978-3-031-81089-3_6.D’yakonova, O.G., (2024). On the Concept of Artificial Intelligence Technologies. Courier of the Kutafin University (MSAL), 3, pp. 52–63. (In Russ.).D’yakonova, O.G., (2024). On the Concept of Artificial Intelligence Technologies. Courier of the Kutafin University (MSAL), 3, pp. 52–63. (In Russ.).Fernández-Llorca, D. et al., (2025). An interdisciplinary account of the terminological choices by EU policymakers ahead of the final agreement on the AI Act: AI system, general purpose AI system, foundation model, and generative AI. Artificial Intelligence and Law, 33, pp. 875–888, doi: 10.1007/s10506-024-09412-y.Fernández-Llorca, D. et al., (2025). An interdisciplinary account of the terminological choices by EU policymakers ahead of the final agreement on the AI Act: AI system, general purpose AI system, foundation model, and generative AI. Artificial Intelligence and Law, 33, pp. 875–888, doi: 10.1007/s10506-024-09412-y.Finocchiaro, G., (2024). The regulation of artificial intelligence. AI & Society, 39, pp. 1961–1968, doi: 10.1007/s00146-023-01650-z.Finocchiaro, G., (2024). The regulation of artificial intelligence. AI & Society, 39, pp. 1961–1968, doi: 10.1007/s00146-023-01650-z.Gstrein, O.J., Haleem, N. and Zwitter, A., (2024). General-purpose AI regulation and the European Union AI Act. Internet Policy Review, 13(3), doi: 10.14763/2024.3.1790.Gstrein, O.J., Haleem, N. and Zwitter, A., (2024). General-purpose AI regulation and the European Union AI Act. Internet Policy Review, 13(3), doi: 10.14763/2024.3.1790.Hacker, P. et al., (2025). The Oxford Handbook of the Foundations and Regulation of Generative AI. Oxford University Press. DOI: 10. 1093/oxfordhb/9780198940272.001.0001.Hacker, P. et al., (2025). The Oxford Handbook of the Foundations and Regulation of Generative AI. Oxford University Press. DOI: 10. 1093/oxfordhb/9780198940272.001.0001.Hacker, P., (2023). AI Regulation in Europe: From the AI Act to Future Regulatory Challenges. Available at: https://arxiv.org/abs/2310.04072 [Accessed 15.08.2025].Hacker, P., (2023). AI Regulation in Europe: From the AI Act to Future Regulatory Challenges. Available at: https://arxiv.org/abs/2310.04072 [Accessed 15.08.2025].Heim, L. and Koessler, L., (2024). Training Compute Thresholds: Features and Functions in AI Regulation. Available at: https://arxiv.org/abs/2405.10799 [Accessed 15.08.2025].Heim, L. and Koessler, L., (2024). Training Compute Thresholds: Features and Functions in AI Regulation. Available at: https://arxiv.org/abs/2405.10799 [Accessed 15.08.2025].Ivanova, E.S., (2022). On the Legal Regulation of Artificial Intelligence in the Russian Federation (A Comparative Legal Analysis). State and Law, 11, pp. 202–206. (In Russ.).Ivanova, E.S., (2022). On the Legal Regulation of Artificial Intelligence in the Russian Federation (A Comparative Legal Analysis). State and Law, 11, pp. 202–206. (In Russ.).Khabrieva, T.Ya., (2025). Legal Problems of Artificial Intelligence Identification. Bulletin of the Russian Academy of Sciences, 7, p. 610, doi: 10.31857/S0869587324070015. (In Russ.).Khabrieva, T.Ya., (2025). Legal Problems of Artificial Intelligence Identification. Bulletin of the Russian Academy of Sciences, 7, p. 610, doi: 10.31857/S0869587324070015. (In Russ.).Kiselev, A.S., (2024). Problems and Prospects of Legal Regulation of Social Relations Related to the Use of Neural Networks. Lex Russica, 2, pp. 140–151, doi: 10.17803/1729-5920.2024.207.2.140-151. (In Russ.).Kiselev, A.S., (2024). Problems and Prospects of Legal Regulation of Social Relations Related to the Use of Neural Networks. Lex Russica, 2, pp. 140–151, doi: 10.17803/1729-5920.2024.207.2.140-151. (In Russ.).Koivisto, I., Koulu, R. and Larsson, S., (2024). User accounts: How technological concepts permeate public law through the EU’s AI Act. Maastricht Journal of European and Comparative Law, 31(3), pp. 412– 432, doi: 10.1177/1023263X241248469.Koivisto, I., Koulu, R. and Larsson, S., (2024). User accounts: How technological concepts permeate public law through the EU’s AI Act. Maastricht Journal of European and Comparative Law, 31(3), pp. 412– 432, doi: 10.1177/1023263X241248469.Kovaleva, N.N. and Zhirnova, N.A., (2024). Issues of Ensuring Personal Data Confidentiality in the Use of Artificial Intelligence Systems. Journal of Russian Law, 28(7), pp. 109–121, doi: 10.61205/S160565900027561-1. (In Russ.).Kovaleva, N.N. and Zhirnova, N.A., (2024). Issues of Ensuring Personal Data Confidentiality in the Use of Artificial Intelligence Systems. Journal of Russian Law, 28(7), pp. 109–121, doi: 10.61205/S160565900027561-1. (In Russ.).Placani, A., (2024). Anthropomorphism in AI: hype and fallacy. AI and Ethics, 4, pp. 691–698, doi: 10.1007/s43681-024-00419-4.Placani, A., (2024). Anthropomorphism in AI: hype and fallacy. AI and Ethics, 4, pp. 691–698, doi: 10.1007/s43681-024-00419-4.Polyakova, T.A. and Smirnov, A.A., (2024). Legal Aspects of Ensuring Information Security in the Use of Artificial Intelligence Technologies. Moscow University Bulletin, 26(3), pp. 89–106, doi: 10.55959/MSU2413-631X-27-15-3-07. (In Russ.).Polyakova, T.A. and Smirnov, A.A., (2024). Legal Aspects of Ensuring Information Security in the Use of Artificial Intelligence Technologies. Moscow University Bulletin, 26(3), pp. 89–106, doi: 10.55959/MSU2413-631X-27-15-3-07. (In Russ.).Price, C., (2025). Algorithmic Advancement in Artificial Intelligence. Available at: https://www.rand.org/pubs/research_reports/RRA3485-1.html [Accessed 15.08.2025].Price, C., (2025). Algorithmic Advancement in Artificial Intelligence. Available at: https://www.rand.org/pubs/research_reports/RRA3485-1.html [Accessed 15.08.2025].Raman, D. et al., (2025). Intolerable Risk Threshold Recommendations for Artificial Intelligence. Available at: https://arxiv.org/abs/2503.05812 [Accessed 15.08.2025].Raman, D. et al., (2025). Intolerable Risk Threshold Recommendations for Artificial Intelligence. Available at: https://arxiv.org/abs/2503.05812 [Accessed 15.08.2025].Sastry, G. et al., (2024). Computing Power and the Governance of Artificial Intelligence. Available at: https://arxiv.org/abs/2402.08797 [Accessed 15.08.2025].Sastry, G. et al., (2024). Computing Power and the Governance of Artificial Intelligence. Available at: https://arxiv.org/abs/2402.08797 [Accessed 15.08.2025].Siddiqui, S. et al., (2025). Promising Topics for U.S. — China Dialogues on AI Safety and Governance. Available at: https://www.oxfordmartin.ox.ac.uk/publications/promising-topics-for-us-chinadialogues-on-ai-safety-and-governance [Accessed 15.08.2025].Siddiqui, S. et al., (2025). Promising Topics for U.S. — China Dialogues on AI Safety and Governance. Available at: https://www.oxfordmartin.ox.ac.uk/publications/promising-topics-for-us-chinadialogues-on-ai-safety-and-governance [Accessed 15.08.2025].Smuha, N.A. and Yeung, K., (2025). The European Union’s AI Act: Beyond Motherhood and Apple Pie? In N.A. Smuha (ed.), The Cambridge Handbook of the Law, Ethics and Policy of Artificial Intelligence. Cambridge University Press, pp. 228–258, doi: 10.1017/9781009367783.015.Smuha, N.A. and Yeung, K., (2025). The European Union’s AI Act: Beyond Motherhood and Apple Pie? In N.A. Smuha (ed.), The Cambridge Handbook of the Law, Ethics and Policy of Artificial Intelligence. Cambridge University Press, pp. 228–258, doi: 10.1017/9781009367783.015.Villalobos, P. et al., (2022). Machine Learning Model Sizes and the Parameter Gap. Available at: https://arxiv.org/abs/2207.02852 [Accessed 15.08.2025].Villalobos, P. et al., (2022). Machine Learning Model Sizes and the Parameter Gap. Available at: https://arxiv.org/abs/2207.02852 [Accessed 15.08.2025].Winter, C. and Bullock, C., (2024). The Governance Misspecification Problem. Institute for Law & AI Working Paper Series, 3, pp. 1–29. Available at: https://law-ai.org/the-governance-misspecification-problem [Accessed 20.09.2025].Winter, C. and Bullock, C., (2024). The Governance Misspecification Problem. Institute for Law & AI Working Paper Series, 3, pp. 1–29. Available at: https://law-ai.org/the-governance-misspecification-problem [Accessed 20.09.2025].Zenin, S.S., Yapryntsev, I.M. and Ivanova, L.V., (2024). The Category of “Legal Responsibility” in the Context of Its Application to Artificial Intelligence Systems. State and Law, 11, pp. 163–173. (In Russ.).Zenin, S.S., Yapryntsev, I.M. and Ivanova, L.V., (2024). The Category of “Legal Responsibility” in the Context of Its Application to Artificial Intelligence Systems. State and Law, 11, pp. 163–173. (In Russ.).Zykov, S.V., (2024). The Problem of Artificial Intelligence Personhood. Actual Problems of Russian Law, 2, pp. 131–139, doi: 10.17803/1994-1471.2024.159.2.131-139. (In Russ.).Zykov, S.V., (2024). The Problem of Artificial Intelligence Personhood. Actual Problems of Russian Law, 2, pp. 131–139, doi: 10.17803/1994-1471.2024.159.2.131-139. (In Russ.).

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