Competition of regulatory paradigms in the field of AI: Liberal and state-centric models and their promotion in the OECD and ITU

Full Text

Introduction

The struggle for leadership in artificial intelligence has become one of the main elements of the US—China rivalry. It unfolds in two dimensions: the development of technological solutions and the formulation of new norms and rules in global governance. This conflict is rooted in divergent approaches to the international regulation of artificial intelligence (AI). They are reflected, on the one hand, in a liberal model based on open-market principles and legal regulation which the United States and the European Union (EU) support and, on the other hand, in a state-centric model focused on sovereignty which the People’s Republic of China promotes (Soriano Gatica, 2025). However, the liberal model does not appear to be monolithic, as it is divided between the EU’s regulatory approach and the US’s market-oriented approach. The importance of international regulation in this field is also confirmed by G. Allison who compares AI risks with Cold War risks linked to the absence of regulation in the nuclear sphere (Allison, 2020).

The struggle to shape norms takes place within key international institutions, where each bloc seeks to universalize its own approach. The Organisation for Economic Co-operation and Development (OECD) has become a pillar of the liberal approach to AI governance and a platform where a synthesis of the American and European models is taking shape. The International Telecommunication Union (ITU) in turn serves as an arena for competition between the liberal and state-centric models and as the main space for advancing the Chinese discourse.

This study focuses on the contestation of norms and rules in international AI regulation, as well as on the mechanisms of their promotion and transformation. It relies on a discourse analysis of proposals and documents from the United States, the European Union and China in order to demonstrate the differences between the three main models of AI governance. The study then examines the process of synthesis and hybridization of these models within the OECD and the ITU.

The theoretical basis of the study is based on regime theory and structural power theory. Regime theory suggests that the core elements of a regime include norms, principles, rules and decision-making procedures (Krasner, 1982). This approach provides tools for analyzing the formation of a fragmented “regime complex” for AI and the existing problems of international regulation (Roberts et al., 2024). Structural power theory, in turn, makes it possible to view the rivalry between the United States and China not as a dispute over rules, but as competition between two “technopoles” for control over “global dependency structures in the digital domain” (Mayer & Lu, 2025).

To achieve the research objective, the authors formed a purposive sample of official documents from the United States, the European Union and the PRC, as well as OECD and ITU materials for 2017–2025. The selected materials cover three substantive blocks:

1) strategic documents in the field of AI,

2) legal and regulatory acts,

3) international initiatives and reports.

Based on the discourse analysis of each body of documents, the authors identified key concepts that reflect the ideological foundations and regulatory priorities of the corresponding models of AI regulation. At the stage of forming the empirical basis of the study, the authors manually selected and verified the keywords identified in legal and regulatory documents in terms of their semantic relevance to the research problem. The resulting data are presented in tables and interpreted through comparative analysis. This made it possible to identify points of divergence and processes of hybridization of the liberal and state-centric approaches in the international institutions under consideration.

Approaches to International Regulation in the Field of Artificial Intelligence

The intensifying technological rivalry between the United States and China is increasingly viewed not merely as an innovation race (Schmid et al., 2025), but also as a broader contest for global leadership or an “AI Cold War” (He & Heeks, 2025). This process is accompanied by a clash of values between Western and Eastern civilizations (McInerney, 2024). These values form the basis of the regulatory paradigms promoted in the field of AI. In this context, the liberal “human-centric” model stands in opposition to the “state-centric” model that places state sovereignty at the center (Arora et al., 2025).

The liberal paradigm aimed at supporting the use of innovative and trustworthy AI rests on democratic values and the idea of protecting human rights (Roberts et al., 2024). Although the United States and China may share common objectives in international platforms, their approaches to specific issues of AI regulation differ. Thus, the United States emphasizes the dominance of the market-based approach, while the European Union by contrast supports the principle of common normative regulation (Sayankina et al., 2025). Both the United States and the EU have chosen the OECD as the main platform for promoting the liberal model of AI regulation.

The state-centric model of AI governance, promoted by China in turn places digital sovereignty above individual freedoms and human rights as understood in the West. This approach is aimed, on the one hand, at restrictive regulation of content and data and, on the other hand, at stimulating industrial and technological development (Ma, 2024).

Against the background of intensifying technological competition between IT powers, the Chinese government views “standardized global governance” as the key to a historical transition from passive adherence to external norms to active creation of global rules. However, at present, such a scenario seems possible only after China overcomes the low internationalization of its own technical regulations and achieves innovative breakthroughs in the field of digital standards (Dong, 2023). The PRC uses the ITU to legitimize its model and promote technological standards that ensure top-down control (Hoffmann, Lazanski & Taylor, 2020).

Thus, China proposes a multilateral approach to the international regulation of AI in opposition to the multistakeholder approach. A similar division could already be observed in the information space (Degterev, Ramich & Piskunov, 2021; Ramich & Piskunov, 2022). In the contemporary world the growing fragmentation of global cyberspace regulation is already visible due to divergences between networked developed countries and networked developing states in their understanding of sovereignty, cybersecurity and the role of international law. This prevents the achievement of consensus (Tang & Li, 2022).

Liberal Model of International AI Regulation

As noted above, the United States and the European Union are the primary proponents of the liberal model of AI regulation, grounded in the primacy of human rights and democratic values. Despite this, their approaches differ to a certain extent in practice. The United States relies on a market-based bottom-up approach to standard-setting, while the European Union gives preference to a regulatory top-down approach (Shahin, 2024, p. 1122). This section presents an analysis of documents to demonstrate the practical implementation of two different approaches serving a single goal.

The European Union’s Regulatory Model

Given its lagging behind the United States in terms of technology, the European Union prioritizes the formation of legal norms. The primary task of this approach is to protect EU sovereignty in AI by establishing a global standard focused on human rights protection and promoting it globally (Mügge, 2024). Discourse analysis of EU documents makes it possible to identify three stages in the development of this strategy.

During 2018–2019, the EU established the foundation of its AI discourse. The data show that “innovation” with 74 mentions and “ethics” with 116 mentions in this period became the most prominent frames (Table 1).

Table 1. Content Analysis of EU Documents, 2018–2025

Keyword	Year
	2018	2019	2020	2021*	2021**	2025***	2025****
Ethical	25	91	9	15	18	6	1
Human rights	1	12	0	0	1	0	0
Innovation	65	9	25	16	97	44	36
Legal framework	5	2	5	10	5	1	1
Rule of law	0	9	1	2	1	0	0
Technological (digital) sovereignty	0	0	1	1	1	6	4
Trustworthy	2	127	9	21	58	5	7

Notes. * — The AI Act; ** — Annexes to the  Communication from the European Commission Fostering a European Approach to AI; *** — AI Continent Action Plan; **** — Apply AI Strategy.
Source: compiled by V.O. Potemkina based on: Coordinated Plan on Artificial Intelligence // European Union. 2018. URL: https://eur-lex.europa.eu/legal-content/EN/ALL/?uri=COM:2018:795:FIN (accessed: 12.11.2025); High-Level Expert Group on Artificial Intelligence // European Commission. 2019. URL: https://www.aepd.es/sites/default/files/2019-12/ai-ethics-guidelines.pdf (accessed: 12.11.2025); White Paper on Artificial Intelligence — A European Approach to Excellence and Trust // European Commission. February 19, 2020. URL: https://commission.europa.eu/system/files/2020-02/commission-white-paper-artificial-intelligence-feb2020_en.pdf (accessed: 12.11.2025); The AI Act // European Commission. 2021. URL: https://artificialintelligenceact.eu/wp-content/uploads/2021/08/The-AI-Act.pdf (accessed: 12.11.2025); Annexes to the  Communication from the Commission to the European Parliament, the European Council, the Council, the European Economic and Social Committee and the Committee of the Regions Fostering a European Approach to Artificial Intelligence // European Commission. April 21, 2021. URL: https://www.eciia.eu/wp-content/uploads/2021/05/1_en_annexe_autre_acte_part1_v8_vf_C4B261EB-ABA4-5C30-1555482869410384_75787.pdf (accessed: 12.11.2025); AI Continent Action Plan // European Commission. April 9, 2025. URL: https://digital-strategy.ec.europa.eu/en/library/ai-continent-action-plan (accessed: 12.11.2025); Apply AI Strategy // European Commission. 2025. URL: https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:52025DC0723 (accessed: 12.11.2025).

These two concepts formed the basis of “trustworthy AI,” which established three core principles: lawful, ethical, and robust. Within this framework, the EU treats the rule of law and human rights as integral components.

In 2020–2021 legal norms began to take shape on the basis of these ethical principles. The 2020 White Paper outlined the EU’s position on technological sovereignty.^[1] The AI Act formalized the concept of trustworthiness into a concrete risk-based approach and made it an element of legal controllability and transparency.^[2] This confirms the earlier argument about the EU’s positioning as a global rule-maker relying on the so-called “Brussels effect” understood as the EU’s ability to extend the norms, regulatory measures and standards of its internal market beyond the Union and turn them into de facto global standards (Bradford, 2020, p. 14).

By 2025, the EU discourse had shifted toward concrete objectives for strengthening “technological sovereignty” with 10 mentions in two documents. After establishing the normative foundation the EU moved to specific infrastructural mechanisms such as AI Factories, the Cloud and AI Development Act and the Data Union. These mechanisms make it possible to control platforms and build the capacity for self-sufficiency.

Market-Based Model for Preserving US Technological Leadership

For the United States, the main objective is to preserve leadership in the field of AI. The United States promotes its AI standards as the basis of global regulation while using international coalitions and simultaneously investing in R&D in order to win global competition (Joshi, 2025, p. 23). Possessing a technological advantage that neither the EU nor the PRC currently has, the United States has a temporary “window of opportunity” to promote its values and standards at the global level. The US strategy has similarly gone through several stages of development. Its distinctive feature was its orientation toward the technical and market aspects of AI regulation.

During 2019–2021, under President Donald Trump, the United States emphasized “leadership” with 7 mentions and “privacy” with 19 mentions in 2019–2020 (Table 2). On the one hand, this highlights the US’s aspiration to preserve global leadership and, on the other hand, demonstrates its commitment to open-market principles. The issues of trust in the US model are centered on technical reliability and flexible normative regulation. The 2021 final report of the National Security Commission on Artificial Intelligence (NSCAI)^[3] identifies the OECD and the ITU as platforms for promoting the US approach in the international arena. The OECD serves as a channel for exporting US norms, while the ITU should ensure the protection of standards and the alignment of technologies with democratic values by preventing the influence of the Chinese approach.

In 2022–2024, during Joseph Biden’s presidential term, the conceptual framework of the US global strategy was established. As expected, the US discourse shifted toward technical categories. During this period the following terms reached peak frequencies: “privacy” with 86 mentions in 2022, “risk management” with 79 mentions in 2023, “trustworthy” with 23 mentions in 2023 and “secure” with 37 mentions in 2022–2023 (Table 2). Documents from this period, such as the NIST AI RMF 1.0^[4] and the Blueprint for an AI Bill of Rights,^[5] establish a system of voluntary and measurable risk-management mechanisms. Unlike the EU AI Act, this system can easily be transformed into a universal standard for global consolidation without the need to adopt binding laws.

 Table 2. Content Analysis of US Documents, 2019–2025

Keyword	Year
	2019	2020	2022	2023	2025
Accountability	0	0	7	15	0
Human rights	0	0	9	1	0
Innovation	4	2	14	2	29
Leadership	6	1	0	4	7
Privacy	11	8	86	31	3
Risk management	0	0	4	79	0
Safety	3	3	25	12	1
Trustworthy	1	1	8	23	6

Source: compiled by V.O. Potemkina based on: Executive Order 13859 of February 11, 2019 — Maintaining American Leadership in Artificial Intelligence // Federal Register.  February 14, 2019. URL: https://www.govinfo.gov/content/pkg/FR-2019-02-14/pdf/2019-02544.pdf (accessed: 12.11.2025); Executive Order of December 3, 2020 — Promoting the Use of Trustworthy Artificial Intelligence in the Federal Government // Federal Register. December 8, 2020. URL: https://www.govinfo.gov/content/pkg/FR-2020-12-08/pdf/2020-27065.pdf (accessed: 12.11.2025); Blueprint for an AI Bill of Rights // The White House. 2022. URL: https://bidenwhitehouse.archives.gov/ostp/ai-bill-of-rights/ (accessed: 12.11.2025); Artificial Intelligence Risk Management Framework (AI RMF 1.0) // NIST. 2023. URL: https://nvlpubs.nist.gov/nistpubs/ai/NIST.AI.100-1.pdf (accessed: 12.11.2025); Winning the Race America’s AI Action Plan // The White House. July 2025. URL: https://www.whitehouse.gov/wp-content/uploads/2025/07/Americas-AI-Action-Plan.pdf (accessed: 12.11.2025).

In 2025, after D. Trump returned to the White House, the United States reaffirmed its commitment to maintaining its status as the architect of global AI infrastructure by engaging partners in its own standards. The techno-economic alliances created by the United States are intended to contain the technological potential of the PRC and maintain US global dominance in AI.

State-Centric Model of AI Regulation in the PRC

As noted above, the PRC’s state-centric model, unlike the liberal model, places digital sovereignty and national security at the center, rather than individual freedoms. This logic is shaped by the need to ensure social stability, national security and economic development (Arora et al., 2025). For this reason, the PRC simultaneously uses strict regulation of digital content while also stimulating technological development (Ma, 2024). Therefore, two foundations of the Chinese strategy can be identified: digital sovereignty for ensuring national security and technological innovation for global leadership. At the same time standardization becomes a key instrument of Chinese “soft power” which allows China to transmit its own regulatory principles in international institutions thereby strengthening its voice and expanding its influence (Dong, 2023, p. 91).

China structures its discourse on digital sovereignty around the concepts of “control,” “national security,” “data security” and “social stability” (Table 3). The publication of the 2021 White Paper^[6] became an important stage in the institutional consolidation of the PRC’s state-centric model. It focused on governing and regulating AI systems to guarantee social stability. The logical continuation of this process was the interim document on the governance of generative artificial intelligence services in 2023,^[7] which consolidated state control over data, models and infrastructure already at the stage of standard-setting. The document is based on socialist values and principles of controllability, with security interpreted as a form of political stability.

Table 3. Content Analysis of PRC Documents, 2017–2025

Keyword	Year
	2017	2021	2022*	2022**	2025
National security	8	1	1	0	1
Social stability	0	1	1	1	0
Control (controllability)	30	6	7	4	3
Data security	1	1	9	2	0
Innovation	63	5	34	1	13

Notes. * — Artificial Intelligence White Paper; ** — Position Paper of the People’s Republic of China on Strengthening Ethical Governance of AI.

Source: compiled by V.O. Potemkina based on: Artificial Intelligence Development Plan // Foundation for Law & International Relations. July 8, 2017. URL: https://flia.org/notice-state-council-issuing-new-generation-artificial-intelligence-development-plan/ (accessed: 12.11.2025); White Paper on Trustworthy A Next Generation Artificial Intelligence Development Plan // Center for Security and Emerging Technologies. September 14, 2021. URL: https://cset.georgetown.edu/publication/white-paper-on-trustworthy-artificial-intelligence/ (accessed: 12.11.2025); Artificial Intelligence White Paper (2022) // Center for Security and Emerging Technologies. June 16, 2022. URL: https://cset.georgetown.edu/publication/artificial-intelligence-white-paper-2022/ (accessed: 12.11.2025); Position Paper of the People’s Republic of China on Strengthening Ethical Governance of Artificial Intelligence (AI) // Ministry of Foreign Affairs, People’s Republic of China. November 17, 2022. URL: https://www.fmprc.gov.cn/eng/zy/wjzc/202405/t20240531_11367525.html (accessed: 12.11.2025); Opinions of the State Council on Deepening the Implementation of the “Artificial Intelligence+” Initiative // Center for Security and Emerging Technologies. September 24, 2025. URL: https://cset.georgetown.edu/publication/china-ai-plus-opinions-2025/ (accessed: 12.11.2025).

The second important foundation for global leadership is technological “innovation.” Its significance is confirmed by the highest frequency of mentions in 2017 with 63 mentions, in 2022 with 34 mentions and in 2025 with 13 mentions (see Table 3). The orientation toward technological progress was fixed in the 2022 White Paper,^[8] where the key vector was the achievement of innovative progress in connection with the engineering component and security, as well as the harmonization of Chinese and global technological standards.

The initial stage of the international promotion of norms was China’s proposal of its own set of ethical principles combining the UN-adopted rhetoric of “AI for Good” and Chinese strategic components such as a “community with a shared future for mankind.”^[9] For the PRC, the key distinction was the promotion of a multilateral governance model, in which states play the decisive role in opposition to multistakeholderism.

In 2025, the PRC consolidated its course toward joint and mutually beneficial regulation of technologies by promoting the idea of AI as an international public good and emphasizing the need to develop global rules on international platforms.^[10]

Competition between Approaches in International Organizations

Organisation for Economic Co-operation and Development

The OECD has become the main organization for combining the US and EU approaches into a single liberal paradigm. This process is characterized primarily by a synthesis of the EU’s legal and ethical approach and US technological standards.

In 2019 the OECD developed AI principles that feature “human rights” with 12 mentions and “the rule of law” with 3 mentions associated with the EU model as well as “risk management” with 8 mentions and “accountability” with 8 mentions associated with the US model (Table 4).

Table 4. Content Analysis of OECD Documents, 2019–2025

Keyword	Year
	2019	2023	2025*	2025**	2025***
Security	11	21	8	3	66
Rule of law	3	1	3	1	14
Control	0	7	0	0	50
Multilateral	0	0	0	4	2
Multistakeholders	0	0	7	2	2
Accountability	8	2	3	8	132
Human rights	12	10	9	9	26
Legal framework	0	1	0	3	11
Development	8	50	24	37	198
Regulation	1	54	0	12	83
Technological sovereignty	0	0	0	0	1
Risk management	8	17	3	15	57
Stability	0	1	0	0	0
SDGs	1	0	0	1	3
Ethical	0	5	2	9	91

Notes. * — Recommendation of the Council on AI; ** — Steering AI’s Future: Strategies for Anticipatory Governance; *** — Governing with AI.
Source: compiled by V.O. Potemkina based on: OECD AI Principles Overview // OECD. 2019. URL: https://oecd.ai/en/ai-principles (accessed: 12.11.2025); Stocktaking for the Development of an AI Incident Definition // OECD Artificial Intelligence Papers. 2023 (October). No. 4. URL: https://www.oecd.org/content/dam/oecd/en/publications/reports/2023/10/stocktaking-for-the-development-of-an-ai-incident-definition_64c69a10/c323ac71-en.pdf (accessed: 12.11.2025); Recommendation of the Council on Artificial Intelligence // OECD. 2025. URL: https://legalinstruments.oecd.org/api/print?ids=648 (accessed: 12.11.2025); Steering AI’s Future: Strategies for Anticipatory Governance // OECD. February 7, 2025. URL: https://www.oecd.org/en/publications/steering-ai-s-future_5480ff0a-en.html (accessed: 12.11.2025); Governing with Artificial Intelligence — The State of Play and Way Forward in Core Government Functions // OECD. September 2025.URL: https://www.oecd.org/content/dam/oecd/en/publications/reports/2025/06/governing-with-artificial-intelligence_398fa287/795de142-en.pdf (accessed: 12.11.2025).

In parallel with changes in EU and US rulemaking in the field of AI during 2020–2023, OECD reports indicate a shift from ethics to direct governance of technology. In 2023, the influence of the US approach became more visible since the OECD document devoted to developing unified terminology in the context of anticipating and monitoring AI threats is dominated by American categories such as “risk management” and “security.” These categories imply the unification of procedures for managing risks and safety while preserving a voluntary technical regime without the strict legal regulation typical of the EU. Meanwhile, at the value level, the document preserves the EU approach where technological risks are perceived through the lens of fundamental values and human rights.

By 2025, the OECD had developed its own procedural model for implementing principles based on the strategy of “anticipatory governance.” This model includes elements of inclusiveness, human rights, accountability, security, flexible regulation, stakeholder engagement and the participation of such structures as the United Nations, the G7, the G20, the EU and the North Atlantic Treaty Organization (NATO).^[11] The main goal of the discourse became the promotion of a normative approach in which innovation is combined with ethical principles and regulatory mechanisms.^[12] This represents a combination of the key conceptual frameworks of the US and EU models.

International Telecommunication Union

The ITU, operating under the auspices of the United Nations, serves as the main arena for competition between the liberal and state-centric models of AI governance. This is due to the fact that the ITU serves as one of the main platforms for shaping the core of the global technology governance regime through the concept of “AI for Good,” which embeds AI into the global initiative aimed at achieving the UN Sustainable Development Goals. China’s deliberate integration into UN-led multilateral formats helps it use the ITU as a space for consolidating the positions of “networked developing countries” in an attempt to assume the role of their representative and promote an alternative discourse in global AI governance (Tang & Li, 2022).

An analysis of UN reports for 2022–2024 reveals that, at the level of terminology, the discourse shifted toward the liberal model. This is confirmed by the high frequency of mentions of terms such as “human rights,” “multistakeholders,” “ethical” and others (Table 5). On the one hand, the ITU consolidates the trend of developing AI technologies as a “public good” while avoiding open geopolitical confrontation. On the other hand, it strengthens the role of the United States and the European Union as leading actors in setting the rules of digital policy through the use of their terminology.

In the 2025 report,^[13] American rhetoric clearly dominates the discussion. The focus shifts toward universal, measurable, and standardized governance of technologies, while preserving guarantees of safety, control, accountability, and risk management. The use of terms such as “safety/security” (135 times), “control” (43 times), and “risk management” (9 times) prevails (see Table 5).

However, the 2025 annual report on AI governance^[14] reveals the hybridization of all three models. It regulates control over AI through infrastructure in line with the US logic. It emphasizes issues of digital inequality and social justice in line with the EU logic. It also introduces the term “sovereign AI” and increases the significance of multilateralism in line with the PRC logic. Turning to statistics, it should be noted that over the past decade, the PRC’s activity in the ITU has grown rapidly: the number of representatives has increased many times over and the share of contributions to the Telecommunication Standardization Sector has exceeded 50%. This creates favorable conditions for the PRC to further promote the Chinese state-centric model of technology regulation at the global level.^[15]

Table 5. Content Analysis of ITU Documents, 2022–2025

Keyword	Year
	2022	2023	2024	2025*	2025**
Security	82	132	191	135	24
Rule of law	5	8	20	0	3
AI for Good	28	31	34	26	26
Control	25	21	38	43	29
Multilateral	4	10	23	7	12
Multistakeholders	27	35	49	15	16
Accountability	9	6	13	20	24
Human rights	205	228	259	7	16
Legal framework	0	1	5	2	0
Development	687	890	1293	128	94
Regulation	30	49	65	42	20
(AI, digital, tech, data) Sovereignty	0	0	0	2	10
Risk management	5	5	10	9	11
Stability	4	8	6	4	1
SDGs	1049	1539	2131	0	1
Ethical	56	100	159	42	28

Notes. * — AI Standards for Global Impact: From Governance to Action; ** — The Annual AI Governance Report 2025.
Source: compiled by V.O. Potemkina based on: United Nations Activities on Artificial Intelligence (AI) // ITU. 2022. URL: https://demo.ifgict.org/wp-content/uploads/2024/09/ITU- UN-AI-Activities-2022-IFGICT.pdf (accessed: 12.11.2025); United Nations Activities on Artificial Intelligence (AI) // ITU. 2023. URL: https://www.itu.int/dms_pub/itu-s/opb/gen/ S-GEN-UNACT-2023-PDF-E.pdf (accessed: 12.11.2025); UN Activities on AI Report 2024 // ITU. 2024. URL: https://www.itu.int/hub/publication/s-gen-unact-2024/ (accessed: 12.11.2025); AI Standards for Global Impact: From Governance to Action // ITU. 2025. URL: https://s41721.pcdn.co/wp-content/uploads/2025/09/ai-standards-for-global-impact.pdf (accessed: 12.11.2025); The Annual AI Governance Report 2025: Steering the Future of AI // ITU. 2025. URL: https://mcit.gov.eg/Upcont/Documents/Reports%20and%20Documents_9112025000_Annual_AI_Governance_Report_2025_Steering_Future_of_AI_9112025.pdf (accessed: 12.11.2025).

Thus, the ITU is transforming from a platform where liberal principles dominated into a hybrid platform for AI regulation, where all three competing models coexist.

Conclusion

The analysis demonstrates that as geopolitical competition between the United States and the PRC intensifies, the liberal model is shifting its priorities from the primacy of the ethical approach in the EU and the rhetoric of free regulation in the United States toward a stricter approach focused on technological sovereignty, the rule of law, risk management and security. This suggests that countries in the liberal bloc are partially adopting the tools of the state-centered model.

Current dynamics reflect changes in the global AI governance regime, which is becoming increasingly fragmented. The OECD has become the normative “core” for Western countries by forming a single global approach based on the combination of the American and European models. Under conditions of competition with the PRC, this approach has so far allowed the United States and the European Union to promote their norms and rules in international institutions.

The ITU has become the central arena for competition between the liberal and state-centric models. For the PRC, this platform serves as a major international space for promoting its own model of international AI regulation. Despite the initial dominance of liberal discourse in the organization, the term “sovereign AI” entered the organization’s terminology in 2025 and clearly refers to the Chinese model. This indicates a transition to a higher level of competition for the right to set international standards, which have become an instrument of structural power. Under current conditions, the PRC seeks to consolidate a multilateral model of global governance in order to deprive its competitors among developed countries of the advantages of the dominant multistakeholder model.

The methodological toolkit of regime theory made it possible to confirm the thesis about the fragmentation of the field of global AI governance, where the key problem is the coordination of norms, principles, rules and procedures under conditions of an emerging regime complex. In turn, the application of structural power theory made it possible to define the nature of relations between the key actors as rivalry between technologically advanced powers for control over the global structures of digital dependence. This rivalry extends beyond the process of establishing international rules.

Ultimately, regardless of their ideological foundations, the liberal and state-centric models are moving toward the sovereigntization of the AI sphere and shifting the priorities of global governance from ethical issues toward ensuring national security and national interests. The OECD and the ITU have become the key instruments for the competing blocs. These institutions serve simultaneously to disseminate their own norms and to counter the promotion of competing approaches.

 

^{1 White Paper on Artificial Intelligence — A European Approach to Excellence and Trust // European Commission. February 19, 2020. URL: https://commission.europa.eu/system/files/2020-02/commission-white-paper-artificial-intelligence-feb2020_en.pdf (accessed: 12.11.2025).}

^{2 The AI Act // European Commission. 2021. URL: https://artificialintelligenceact.eu/wp-content/uploads/2021/08/The-AI-Act.pdf (accessed: 12.11.2025).}

^{3 Final Report // National Security Commission of Artificial Intelligence. 2021. URL: https://www.dwt.com/-/media/files/blogs/artificial-intelligence-law-advisor/2021/03/nscai-final-report--2021.pdf (accessed: 12.11.2025).}

^{4 Artificial Intelligence Risk Management Framework (AI RMF 1.0) // NIST. 2023. URL: https://nvlpubs.nist.gov/nistpubs/ai/NIST.AI.100-1.pdf (accessed: 12.11.2025).}

^{5 Blueprint for an AI Bill of Rights // The White House. 2022. URL: https://bidenwhitehouse.archives.gov/ostp/ai-bill-of-rights/ (accessed: 12.11.2025).}

^{6 White Paper on Trustworthy A Next Generation Artificial Intelligence Development Plan // Center for Security and Emerging Technologies. September 14, 2021. URL: https://cset.georgetown.edu/publication/white-paper-on-trustworthy-artificial-intelligence/ (accessed: 12.11.2025).}

^{7 Interim Measures for the Management of Generative Artificial Intelligence Services // Cyrilla. July 10, 2023. URL: https://cyrilla.org/api/files/1728288405530qs2ud3o27ys.pdf (accessed: 12.11.2025).}

^{8 Artificial Intelligence White Paper (2022) // Center for Security and Emerging Technologies. June 16, 2022. URL: https://cset.georgetown.edu/publication/artificial-intelligence-white-paper-2022/ (accessed: 12.11.2025).}

^{9 Position Paper of the People’s Republic of China on Strengthening Ethical Governance of Artificial Intelligence (AI) // Ministry of Foreign Affairs, People’s Republic of China. November 17, 2022. URL: https://www.fmprc.gov.cn/eng/zy/wjzc/202405/t20240531_11367525.html (accessed: 12.11.2025).}

^{10 Opinions of the State Council on Deepening the Implementation of the “Artificial Intelligence+” Initiative // Center for Security and Emerging Technologies. September 24, 2025. URL: https://cset.georgetown.edu/publication/china-ai-plus-opinions-2025/ (accessed: 12.11.2025).}

^{11 Steering AI’s Future: Strategies for Anticipatory Governance // OECD. February 7, 2025. URL: https://www.oecd.org/en/publications/steering-ai-s-future_5480ff0a-en.html (accessed: 12.11.2025).}

^{12 Governing with Artificial Intelligence — The State of Play and Way Forward in Core Government Functions // OECD. September 2025.URL: https://www.oecd.org/content/dam/oecd/en/publications/reports/2025/06/governing-with-artificial-intelligence_398fa287/795de142-en.pdf (accessed: 12.11.2025).}

^{13 AI Standards for Global Impact: From Governance to Action // ITU. 2025. URL: https://s41721.pcdn.co/wp-content/uploads/2025/09/ai-standards-for-global-impact.pdf (accessed: 12.11.2025).}

^{14 The Annual AI Governance Report 2025: Steering the Future of AI // ITU. 2025. URL: https://mcit.gov.eg/Upcont/Documents/Reports%20and%20Documents_9112025000_Annual_AI_Governance_Report_2025_Steering_Future_of_AI_9112025.pdf (accessed: 12.11.2025).}

^{15 NIST Report: A New “Great Game?”: China’s Role in International Standards for Emerging Technologies // Exovera’s Center for Intelligence Research and Analysis. August 31, 2022. P. 18–24. URL: https://cira.exovera.com/research-analysis/china/nist-report-a-new-great-game-chinas-role-in-international-standards-for-emerging-technologies/ (accessed: 12.11.2025).}

About the authors

Mirzet S. Ramich

RUDN University

Author for correspondence.
Email: ramich-ms@rudn.ru
ORCID iD: 0000-0003-1479-2785
SPIN-code: 1830-1087

PhD (Political Science), Assistant Professor, Department of Theory and History of International Relations

2, 10, Miklukho-Maklaya St, Moscow, 117198, Russian Federation

Victoria O. Potemkina

RUDN University

Email: potemkinavik@yandex.ru
ORCID iD: 0009-0001-9004-4685
SPIN-code: 9876-9865

Student, Department of Oriental and African Studies

2, 10, Miklukho-Maklaya St, Moscow, 117198, Russian Federation

Ekaterina V. Demina

RUDN University

Email: demina.ekaterina.v@yandex.ru
ORCID iD: 0009-0002-9568-1381

Assistant Professor, Department of Foreign Languages

2, 10, Miklukho-Maklaya St, Moscow, 117198, Russian Federation

References

Supplementary files