Montreal AI Ethics Institute

Democratizing AI ethics literacy

What is Sovereign Artificial Intelligence?

✍️ By Sun Gyoo Kang1

Disclaimer: The views expressed in this article are solely my own and do not reflect my employer’s opinions, beliefs, or positions. Any opinions or information in this article are based on my experiences and perspectives. Readers are encouraged to form their own opinions and seek additional information as needed.

Introduction

In today’s world, countries around the world fiercely defend their borders, economic competitiveness, data, and natural resources. Now align this concept of national sovereignty to the rapidly evolving world of artificial intelligence. Indeed, this is the main idea behind Sovereign Artificial Intelligence (“Sovereign AI”).2 Sovereign AI refers to a nation’s ability to independently develop, maintain, and deploy AI systems, including the necessary data and infrastructure, with minimal external dependencies.3

But why is Sovereign AI gaining so much attention now?4 As artificial intelligence becomes increasingly central to everything from defence and economic competitiveness to everyday services such as banking, employment and retail, governments are realizing the strategic importance of not being entirely reliant on foreign AI technologies or firms.5 The debate around Sovereign AI raises crucial questions: Is it necessary for national security and prosperity? What are the potential negative sides of countries pursuing their independent AI paths?

This article will explore its historical roots, go through the key arguments both for and against its pursuit, and ultimately aim to provide a clear understanding of this important technological and geopolitical concept.

History of Sovereign AI

While it has recently gained significant traction, Sovereign AI doesn’t come from a vacuum. National interests, although often unstated, were a driving force behind early technology development during the Cold War, including the development of artificial intelligence.6 Many countries competed to achieve technological superiority. Globalization, combined with the rapid growth of the Internet in the late 20th and early 21st centuries, accelerated the internationalization of AI development. As big tech firms emerged to lead AI development, the field’s focus temporarily shifted from national interests to global commercial innovation and deployment of AI technologies across borders.7

A key turning point in the shift towards Sovereign AI thinking was the growing awareness of data privacy8 and security concerns in the 2010s.9 The high volume of data breaches and misuse, compounded by the growing consolidation of data within a small number of multinational corporations, has fueled a renewed emphasis on national data sovereignty in an era of economic globalization.10 Indeed, regulations such as the European Union’s General Data Protection Regulation (GDPR)11 and the California Consumer Privacy Act (CCPA)12 have sent a strong signal to move towards asserting digital sovereignty and the right of nations to protect their citizens’ data.

Growing geopolitical rivalry between world powers has emerged as the key driver of renewed focus on developing nationally controlled AI systems.13 Countries now treat AI capabilities as vital strategic assets, recognizing their decisive role in economic power, national defence, and military superiority.14 The realization that reliance on foreign AI technologies could create great vulnerabilities, both economically and in terms of security, has driven governments to prioritize the development of local AI ecosystems.15

Positive Arguments for Sovereign AI

Advocates for Sovereign AI argue that it is not merely a desirable goal, but a necessary one for nations in the 21st century. Their arguments often centre on key areas of national interest and advantage:

National Security and Defence: As AI becomes very important for military systems, intelligence operations, and cybersecurity, countries seek independent control to avoid strategic vulnerabilities. Relying on foreign AI systems could compromise defence capabilities and national independence, especially in an unstable geopolitical environment that we’re currently experiencing today.16

Economic Competitiveness and Innovation: For many powerful nations, such as the United States or China, sovereign AI is seen as crucial for economic leadership. Investing domestically in AI will maintain their global competitiveness. Their objective is to develop AI capabilities aligned with their economic priorities, while creating high-skilled jobs and capturing the benefits of AI-driven growth.17 A great example is South Korea’s Basic AI Act, which is a law to foster AI competitiveness.18

Data Privacy and Security: Data privacy and security concerns, as previously discussed, are central to the concept of Sovereign AI.19 Nations seek to safeguard citizen data under their legal frameworks and regulations. Sovereign AI enables nations to establish independent national data governance structures that reflect local values and legal systems.20 This provides control over data flows and mitigates the risk of foreign access or misuse of sensitive information for nations.21

Alignment with National Values and Ethics: Furthermore, and beyond considerations of security and economic advantage, the doctrine of Sovereign AI is predicated on the imperative to align AI development and deployment with a nation’s distinct values and ethical precepts.22 Recognizing the divergence among nations regarding matters such as privacy, bias, fairness, and the societal role of artificial intelligence, Sovereign AI empowers individual states to define and control the ethical and societal ramifications of AI within their jurisdictional boundaries, thereby precluding the imposition of global norms or the prioritization of external national or corporate interests.23

Negative Arguments and Concerns about Sovereign AI

While the arguments for Sovereign AI are compelling to many, the concept also faces significant criticisms and raises valid concerns. Opponents and skeptics point to potential downsides and challenges:

Duplication of Effort and Inefficiency: One major concern is that if every nation strives for complete Sovereign AI, it could lead to a massive duplication of effort and resources globally.24 The substantial costs associated with AI development, encompassing vast data requirements, significant computational resources, and specialized expertise, raise concerns about the efficiency of individual nations pursuing comprehensive, independent AI ecosystems.25 Such an approach risks redundancy and may potentially hinder the overall advancement of the field.

Risk of Protectionism and Reduced Global Collaboration: The pursuit of Sovereign AI carries the inherent risk of fostering protectionist practices within the technology sector.26 Nations may be incentivized to implement restrictive measures concerning cross-border data flows, market access, and preferential treatment of domestic AI enterprises.27 Such actions could impede global collaboration in AI research and development, curtail the free exchange of intellectual capital and skilled personnel, and ultimately slow down the pace of innovation.

Potential for “AI Nationalism” and Geopolitical Tensions: The concept of Sovereign AI is susceptible to conflation with nationalist agendas.28 This intertwining risk exacerbates geopolitical tensions, particularly if nations adopt a zero-sum perspective on AI, leading to aggressive competition for dominance.29 An “AI Arms Race,” characterized by prioritization of military AI capabilities and mutual suspicion, represents a potential adverse consequence of unchecked Sovereign AI ambitions.30

Ethical Concerns and Lack of Global Oversight: The pursuit of Sovereign AI independent from international cooperation risks the fragmentation of AI ethics and governance.31 In the absence of globally harmonized ethical frameworks, inconsistent or even conflicting standards for AI development and deployment may arise across nations.32 This complicates matters such as cross-border data flows, accountability for AI-related harms, and the establishment of universally recognized standards for AI safety.33 Consequently, some posit that AI ethics and governance necessitate a global, rather than solely national or sovereign approach.

Conclusion

Sovereign AI is not an easy topic to grasp. This drive of the nations is motivated by legitimate concerns surrounding national security, economic competitiveness, data privacy, and the desire to align AI development with national values. Proponents argue that it is essential for nations to maintain autonomy and thrive in an AI-driven world.34 Nevertheless, the pursuit of Sovereign AI presents significant drawbacks and risks. Concerns regarding inefficiency, protectionism, geopolitical tensions, and ethical challenges require consideration.35

In conclusion, the optimal path forward likely involves a delicate balance. Complete technological isolation and purely nationalistic approaches to AI are neither feasible nor desirable. Nations should strive for a nuanced “Smart Sovereignty” in AI. This approach necessitates strategic investment in key AI capabilities, combined with active engagement in international collaborations, the sharing of best practices, and concerted efforts to establish common ethical frameworks.36

Footnotes

  1. Law and Ethics in Tech Law and Ethics in Tech | Medium ↩︎
  2. Sovereign AI, https://macquariedatacentres.com/blog/is-sovereign-ai-the-future-of-artificial-intelligence/ (last visited Feb. 4, 2025) ↩︎
  3. Id. ↩︎
  4. GZERO Staff, British PM Wants ‘Sovereign AI’, GZERO MEDIA, https://www.gzeromedia.com/gzero-ai/british-pm-wants-sovereign-ai (last visited Feb 3, 2025). ↩︎
  5. Benjamin Cedric Larsen, The Geopolitics of AI and the Rise of Digital Sovereignty, BROOKINGS, https://www.brookings.edu/articles/the-geopolitics-of-ai-and-the-rise-of-digital-sovereignty/ (last visited Jan 30, 2025). ↩︎
  6. Total Military Insight, Cold War and Technology, TOTAL MILITARY INSIGHT, https://totalmilitaryinsight.com/cold-war-and-technology/ (last visited Jan. 23, 2025). ↩︎
  7. Scott Rosenberg, Who’s winning the AI race, AXIOS (Nov. 27, 2024), https://www.axios.com/2024/11/27/ai-race-openai-google-meta-anthropic (last visited Feb. 1, 2025) ↩︎
  8. Fotis Law Firm, Case Study on Facebook’s Data Breach, FOTIS LAW, https://fotislaw.com/public/lawtify/case-study-on-facebooks-data-breach (last visited Feb. 5, 2025). ↩︎
  9. Karl Manheim & Lyric Kaplan, Artificial Intelligence: Risks to Privacy and Democracy, 21 Yale J.L. & Tech. 106, 106-49 (2019), available at https://yjolt.org/sites/default/files/21_yale_j.l._tech._106_0.pdf (last visited Feb. 2, 2025). ↩︎
  10. Id. ↩︎
  11. General Data Protection Regulation (GDPR), Regulation 2016/679, 2016 O.J. (L 119) 1 (EU). ↩︎
  12. California Consumer Privacy Act (CCPA), Cal. Civ. Code §§ 1798.100–.199 (West 2020). ↩︎
  13. The Geopolitics of AI and the Rise of Digital Sovereignty, supra note 4. ↩︎
  14. Id. ↩︎
  15. World Economic Forum, These 5 countries are leading the global AI race – here’s how they’re doing it, WORLD ECONOMIC FORUM (May 29, 2024), https://www.weforum.org/stories/2024/05/these-5-countries-are-leading-the-global-ai-race-heres-how-theyre-doing-it/ (last visited Dec. 12, 2024) ↩︎
  16. Google Public Policy Team, AI and the Future of National Security, GOOGLE BLOG (Oct. 19, 2023), https://blog.google/technology/safety-security/ai-and-the-future-of-national-security/ (last visited Feb. 4, 2025). ↩︎
  17. Swati Srivastava & Justin Bullock, AI, Global Governance, and Digital Sovereignty, arXiv:2410.17481 (Oct. 24, 2024), https://arxiv.org/abs/2410.17481 (last visited Dec. 10, 2024). ↩︎
  18. Biz Chosun, Korea passes AI Basic Act, second globally, enhancing national AI competitiveness, BIZ.CHOSUN (Dec. 26, 2024), https://biz.chosun.com/en/en-it/2024/12/26/66W2Z3RX6FE7FMPXMR73T26SKY/ (last visited Nov. 3, 2023). ↩︎
  19. Artificial Intelligence: Risks to Privacy and Democracy. Supra note 8. ↩︎
  20. Id. ↩︎
  21. Katharine Miller, Privacy in the AI Era: How Do We Protect Our Personal Information?, STANFORD HAI (Mar. 18, 2024), https://hai.stanford.edu/news/privacy-ai-era-how-do-we-protect-our-personal-information (last visited Jan. 18, 2025). ↩︎
  22. Cathy Roche, P.J. Wall & Dave Lewis, Ethics and diversity in artificial intelligence policies, strategies and initiatives, AI and Ethics, https://doi.org/10.1007/s43681-022-00218-9 (published online Oct. 6, 2022) (last visited Feb. 1, 2025). ↩︎
  23. Didar Zowghi & Muneera Bano, AI for all: Diversity and Inclusion in AI, AI and Ethics, https://doi.org/10.1007/s43681-024-00485-8 (published online May 2, 2024) (last visited Jan. 8, 2025). ↩︎
  24. Ahmed Banafa, The Rise of Sovereign AI: A Technological Race with National Stakes, LINKEDIN (Mar. 19, 2024), https://www.linkedin.com/pulse/rise-sovereign-ai-technological-race-national-stakes-banafa-zxsjc/ (last visited Dec. 10, 2024). ↩︎
  25. Chris Stokel-Walker, Sovereign AI, CODA STORY, https://www.codastory.com/authoritarian-tech/sovereign-ai/ (last visited Feb. 3, 2025). ↩︎
  26. Konstantinos Komaitis, Esteban Ponce de León, Kenton Thibaut, Trisha Ray, and Kevin Klyman, The Sovereignty Trap, ATLANTIC COUNCIL (Jul. 26, 2024), https://www.atlanticcouncil.org/blogs/geotech-cues/the-sovereignty-trap/ (last visited Dec. 2, 2024). ↩︎
  27. Id. ↩︎
  28. Susan Ariel Aaronson, The Age of AI Nationalism and Its Effects, CIGI (Sep. 30, 2024), https://www.cigionline.org/publications/the-age-of-ai-nationalism-and-its-effects/ (last visited Jan. 15, 2025). ↩︎
  29. Id. ↩︎
  30. The Sovereignty Trap, supra note 25 ↩︎
  31. Id. ↩︎
  32. John Letzing, World Economic Forum, What is ‘sovereign AI’ and why is the concept so appealing and fraught?, WORLD ECONOMIC FORUM (Nov. 13, 2024), https://www.weforum.org/stories/2024/11/what-is-sovereign-ai-and-why-is-the-concept-so-appealing-and-fraught/ (last visited Jan. 16, 2025). ↩︎
  33. The Sovereignty Trap, supra note 25 ↩︎
  34. Angie Lee, What Is Sovereign AI?, NVIDIA BLOG (Feb. 28, 2024), https://blogs.nvidia.com/blog/what-is-sovereign-ai/ (last visited Dec. 20, 2024). ↩︎
  35. The Age of AI Nationalism and Its Effects, supra note 27 ↩︎
  36. Muath Alduhishy, World Economic Forum, Sovereign AI: What is it and which ways are states building it?, WORLD ECONOMIC FORUM (Apr. 25, 2024), https://www.weforum.org/stories/2024/04/sovereign-ai-what-is-ways-states-building/ (last visited Dec. 18, 2024). ↩︎
Want quick summaries of the latest research & reporting in AI ethics delivered to your inbox? Subscribe to the AI Ethics Brief. We publish bi-weekly.