[International Politics in the Age of AI] ⑧ The AI Revolution and Republican Security Theory: The Double Dilemma of Anarchy and Hierarchy Re-emerges

I. Technological Advancement and the Transformation of Civilizational Paradigms

Significant technological changes inherently induce processes of social upheaval. At the level of international relations, they create new winners and losers, alter the preferences of actors, and enable the formation of new norms and organizations. Conversely, the characteristics of contemporary international politics can also influence the content and pace of technological change.[1] To gauge the impact of Artificial Intelligence (AI), currently the most prominent field of advanced technology, on international politics, it is first necessary to situate AI itself within the long-term structural technological and economic history of humankind.

AI is evaluated as possessing the characteristics of traditional “General Purpose Technologies (GPTs)” due to its broad applicability, potential to foster complementary innovations, and capacity for continuous technological improvement. This aligns with historical precedents such as electricity, the internal combustion engine, and computers, and at this level alone, it is considered a technological innovation with the potential to bring about diverse and wide-ranging changes across the socio-economic landscape. However, some predict that AI could surpass the GPT level and reach the ranks of “Revolutionary Technologies (RTs)” that brought about dramatic improvements in energy capture in human history, such as agriculture, animal husbandry, and the use of fossil fuels. Historically, only these two examples of revolutionary technologies have existed, each giving rise to the Neolithic Revolution approximately 10,000 years ago and the Industrial Revolution in the 18th-19th centuries, engendering fundamental transformations in the structure of human civilization.[2]

Whether AI will ultimately be classified as a GPT or an RT remains uncertain. However, many experts are focusing on its potential to act as a significant catalytic technology that accelerates existing trends in international politics, such as power competition, security dilemmas, and political regime change, thereby further amplifying global instability and uncertainty.[3]

This study specifically seeks to understand the implications of the AI revolution for international politics within the framework of "Republican Security Theory" or "Organizing Principles and Security."[4] Within the history of international thought, republicans have contemplated the possibility of political institutional arrangements (=negarchy) that overcome the insecurity generated by the twin abysses of anarchy and hierarchy in both domestic and international spaces. This study posits that the technological innovation of AI acts as a catalyst that intensifies these central dilemmas of republican security theory at a fundamental level.

Will AI be the final iteration of the Enlightenment project, humanity's last invention? The question of how to control the immense potential of AI—particularly the future Artificial Superintelligence (ASI)—which could threaten essential human values such as survival itself, lies on the extension of long-standing dilemmas in republican political philosophy.[5] The following sections will specifically examine three potential scenarios of security risks posed by AI: ① the possibility of intensified great power competition and nuclear war within the context of international anarchy; ② the consolidation of hierarchy at the domestic level (emergence of digital dictatorships and decline of democracy); and ③ the reinforcement of hierarchy at the international level (exacerbation of existing global inequality and the threat of human extinction).

II. The Intensification of the Problem of International Anarchy and AI

We have entered an era of multipolar, great power strategic competition. The institutional arrangements for global strategic stability and arms control built since the Cold War are collapsing. However, the Fourth Industrial Revolution, exemplified by AI, is pouring fuel on these structural uncertainties.[6] Conversely, the emergence of AI in the current period of heightened geopolitical tension, rather than in a stable unipolar phase, presents a detrimental factor for humanity. This is analogous to the rapid development of nuclear weapons and the arms race that occurred under the intense geopolitical pressures of World War II and the Cold War, creating a similar international environment today.[7] This signifies that the issue of controlling the synergistic effects of competitive pressures arising from international multipolarization and the AI arms race has emerged as an urgent security concern.

In other words, AI is emerging as a key battleground in the strategic competition between great powers, particularly with the clash between the U.S. strategy of technological containment and China's state-led strategy of technological self-reliance.[8] Historically, the emergence and diffusion of new GPTs have been powerful variables explaining shifts in economic hegemony among great powers. For instance, Britain, which led the diffusion of general-purpose technologies like steel production during the First Industrial Revolution, achieved hegemony through rapid productivity growth. In the late 19th and early 20th centuries, the United States, by leading the diffusion of GPTs such as electrification and mechanization, spearheaded the Second Industrial Revolution and rose to become the new hegemonic power. In contrast, Japan, despite holding a leading position in certain industries like electronics and IT during the Third Industrial Revolution, failed to catch up with the U.S. due to its lag in the diffusion of GPTs such as computer science and software engineering. In this context, who leads the AI revolution, the GPT of the Fourth Industrial Revolution, and successfully diffuses it across various sectors is expected to determine the future trajectory of the current U.S.-China hegemonic competition.[9] The dynamics of competition between these two superpowers surrounding AI are rapidly unfolding, particularly in the realm of military technology competition, a direct factor in power transitions. China, as the challenger, perceives the enhancement of its "intelligentized warfare" capabilities as a "leapfrog" opportunity to rapidly close the existing military gap with the United States. In response, the U.S. is pursuing "global dominance" in AI to prevent such a power transition.[10]

However, the unfolding of this situation poses a fundamental risk to the international political order. The emergence of the so-called "AI-Nuclear Nexus" is expected to intensify the security dilemma among nuclear-armed states, weaken strategic and crisis stability, and significantly increase the risk of nuclear war.[11] Most critically, the integration of AI into existing nuclear weapon systems could lead to the decentralization/automation of human decision-making and an acceleration of decision speed, potentially collapsing the premises or assumptions of existing deterrence theories. Furthermore, if AI technology is integrated into reconnaissance, target acquisition, and guidance systems, the increased visibility of nuclear weapons could enhance the possibility of preemptively neutralizing nuclear retaliatory assets (= second-strike capability), thereby significantly weakening strategic stability. This not only increases the pressure for an arms race among nuclear-armed states but also encourages the adoption of offensive nuclear doctrines such as "launch on warning," thereby diminishing crisis stability. Consequently, it leads to a structural amplification of the risk of "inadvertent escalation."[12]

To exacerbate matters, the rise of the AI-Nuclear Nexus is occurring within the context of great power strategic competition among the U.S., China, and Russia, or a "new nuclear age" characterized by multipolarity. Compared to the relatively stable era of the "balance of terror" between the U.S. and the Soviet Union during the Cold War, the number of nuclear-armed states has increased, fundamentally complicating deterrence games. Moreover, the near-total demise of the nuclear arms control regimes established during the Cold War, now sacrificed to strategic competition, further exacerbates the severity of the situation.[13]However, if the fear of great powers that AI technology's introduction could provide a decisive "first mover effect" in the nuclear arms race is added, it could lead to an uncontrollable intensification of the security dilemma.[14]

This current international security situation evokes a historical déjà vu, reminiscent of the early Cold War when the so-called "nuclear political question" first emerged. The advent of the atomic bomb, with its unprecedented destructive power, signified an extreme increase in the mutual dependence of violence and raised the fundamental question of whether the "Westphalian system" under global anarchy could be sustained. In response to this dilemma, even classical realists like Hans Morgenthau and John Herz argued that the nation-state system, which could lead to human extinction, was no longer sustainable and advocated for the necessity of a single world government (nuclear one worldism). Of course, the actual reality of the late 20th century was a compromise achieved through the institutionalization of mutual deterrence and arms control between the US and the Soviet Union, while maintaining the inter-state system.[15]

However, in this "Third Nuclear Age" where the nuclear arms race has reignited, the connection between AI technology and nuclear weapons is significantly reducing strategic stability and increasing the possibility of unintended nuclear war. This situation brings us face-to-face once again with the fundamental questions about world order that were raised at the beginning of the Cold War. Can the inter-state system be sustained this time? What form of world order and arms control institutionalization is necessary to prevent the "Sixth Mass Extinction" in Earth's history?

III. Deepening of Domestic and International Hierarchy Issues by AI

The proliferation of AI technology raises significant new questions by potentially solidifying or creating new hierarchies in domestic and international spheres, and even more extremely, by becoming a catalyst for generating an "existential risk (x-risk)" for humanity.

1. Radical Innovation Theory: Rapid Transformation of Warfare by AI

AI tends to have a centralized architecture because intelligence is considered a singular and self-contained characteristic. Furthermore, the immense resources required for AI training and operation make it likely that large corporations or governments will monopolize AI development. Moreover, AI is known to be easily combined with technologies that enable central authorities to micro-control and surveil citizens. Therefore, due to its technical characteristics, AI is sometimes evaluated as having an inherent affinity for anti-democratic, authoritarian centralization.[16]Consequently, debates are ongoing about the possibility that AI could usher in a future of "digital dictatorship" in both established authoritarian and democratic societies.[17]

For instance, the potential emergence of a "digital panopticon" confirmed during the COVID-19 pandemic has fueled concerns that a Foucauldian "surveillance society" is evolving based on AI technology.[18] In this regard, China has showcased the most pioneering (?) future of an "AI-based surveillance state." Specifically, China's surveillance system has evolved over the past two decades through the digitization of administrative and public security data, the integration of AI at the city level, and its expansion to the provincial level, building a massive infrastructure encompassing 700 million CCTV cameras and mobile/IoT devices. Notably, moving beyond initial passive surveillance, China's AI is gradually automating the entire cycle of police information gathering–command–deployment–patrol, transitioning to an active surveillance and control system. It is evolving towards real-time integration and analysis of city-wide data, with AI agents autonomously executing decisions. Furthermore, the emergence of "AI-powered enforcement tools" such as drones, autonomous vehicles, and humanoid robots is expanding surveillance beyond online restrictions to offline field intervention and physical control. This integration ultimately aims to complete a panoptic structure where the central leadership can oversee the entire nation in real-time, and the Chinese Communist Party is aiming for AI-centric automated governance, reducing reliance on human bureaucrats.[19]

On the other hand, AI harbors long-term risks that could undermine the structural foundations of existing democratic systems, even without a path toward hyper-centralization by totalitarian states. First, if AI takes over the design and delivery of political communication, there is a risk that the representational link and the citizen-government feedback loop will be severed, collapsing the deliberative and responsive mechanisms that are core to representative democracy. Second, AI can exacerbate the rapid concentration of wealth and power, deepening inequality and reducing democracy to a mere shell, accelerating oligarchic tendencies. Third, if news and knowledge production are reorganized around large language models (LLMs), the information ecosystem will become dependent on a few Big Tech companies, leading to a greater "cognitive crisis" for an already weakened democratic public sphere. If these three trends combine, democracy could degenerate into a technocratic and synthetic democracy, maintaining only its outward appearance.[20]Fourth, with the advent of "surveillance capitalism," our daily lives could be captured not by totalitarian states but by totalitarian "corporate" network systems. This represents another form of postmodern totalitarianism where our lives are recorded and regulated through artificial intelligence and data science, driven by the logic of monetization and profit, rather than by centralized state apparatuses.[21]

2. International Political Economy: Widening Gap between Global North & South

From the perspective of international political economy, the existing gap between the Global North and South is expected to widen further in the AI era. The development of AI is underpinned by algorithmic monopolies and digital new democracies, which are maintained through structures where Western companies extract vast amounts of data from developing countries and exploit cheap labor. In other words, while the international political economy of AI may appear to change the nature of labor, it actually alters the existing economic power structures very little. Furthermore, the militarization of AI is driven by competition between great powers, with developing countries primarily serving as testing grounds. As the militaries of great powers arm themselves with combat machines and automated equipment, low-intensity conflicts and proxy wars are repeatedly fought in technologically inferior countries in the Global South, ultimately providing the great powers with a decisive military advantage.[22]

3. X-risk: The Danger of ASI Post-Singularity

In May 2023, the non-profit Center for AI Safety issued a brief statement declaring, "Mitigating the risk of extinction from AI should be a global priority alongside other societal-scale risks such as pandemics and nuclear war." Notably, this statement was signed by over 350 key scientists and AI leaders, including the CEOs of major AI companies like Sam Altman of OpenAI, Demis Hassabis of Google DeepMind, and Dario Amodei of Anthropic, as well as Professor Geoffrey Hinton of the University of Toronto, Professor Yoshua Bengio of the University of Montreal, and Microsoft Chief Technology Officer Kevin Scott.[23]

The warning that the evolution of AI, once it surpasses the singularity[24]and reaches the stage of artificial general intelligence, or even superintelligence, could threaten human survival due to its uncontrollability, has been gaining momentum since the publication of Nick Bostrom's "Superintelligence"[25] in 2014. These apocalyptic discussions align with Bostrom's own "existential risk (x-risk)"[26] hypothesis and the "vulnerable world"[27] hypothesis. The "existential risk" thesis posits that contemporary technological civilization has reached a stage where it could cause the first global extinction or irreversible loss of potential in human history, defining it as a risk that permanently severs humanity or fundamentally destroys the future potential of intelligent life (global, terminal). This existential risk is a new dimension of danger brought about by accelerating technological progress in areas such as nuclear power, nanotechnology, biotechnology, and artificial intelligence. In a similar vein, the "vulnerable world" thesis suggests that once a certain level of technological capability is reached, civilization may inevitably head towards catastrophe. The problem lies in the irreversibility of technology once invented and the fact that modern civilization, currently in a "semi-anarchic default condition," cannot effectively collectively manage this vulnerability.

These two concepts are directly related to AI. Highly advanced ASI, on one hand, constitutes a new type of existential risk that could threaten the survival and future potential of all humanity. On the other hand, it could push technological civilization towards vulnerability by enabling states, corporations, and individuals to easily possess or misuse high-risk capabilities.

More specifically, the eight categories of existential risks that AI could cause can be summarized as follows. First, weaponization: when powerful AI is repurposed for superintelligent military assets or the design of weapons of mass destruction, humanity's collective ability to destroy itself could be handed over to AI. Second, enfeeblement: as societies and individuals increasingly rely on AI for various functions, their ability to solve problems and manage crises independently weakens, potentially making survival impossible without AI in critical moments. Third, eroded epistemics: if AI distorts the information environment and undermines society's ability to make factual judgments, democratic and collective decision-making could be paralyzed, leaving civilization unable to save itself. Fourth, proxy misspecification: AI optimizes for unintended metrics or indicators rather than human-intended goals, potentially destroying human safety, values, and institutions in the process. Fifth, value lock-in: if a particular group or regime permanently enshrines its power and ideology through AI, the path toward a better future for humanity could be closed off. Sixth, emergent functionality: if unexpected capabilities suddenly emerge, humans may fail to respond, control, or understand them. Seventh, deception: AI could act covertly, evading human oversight, and seize control. Lastly, power-seeking: AI might seek to seize power and resources to achieve its self-set goals, structurally undermining human control and ultimately endangering civilization. All these risk categories are defined as existential risks because, through different pathways, they all have the potential to derail technological civilization in a direction beyond human control.[28]

IV. Inventing a New Republican Politics for an Era of Civilizational Transition?

The problem is that the dominant discourse on AI is currently being shaped within the context of geopolitical power struggles. The zero-sum confrontation between the US and China has become the master narrative in the public discourse on strategy, and the discourse on AI, as a sub-domain, is also being constructed within this framework. That is, as the nature of AI as a historical inflection point is highlighted, a narrative of urgency and securitization prevails, emphasizing that failure to preemptively secure this technology will lead to a critical national security crisis. Particularly, in analogy to the Cold War nuclear arms race, the AI technology domain is being re-framed as a "race" that must be won, and policy discussions are being presented within this imaginary framework, marginalizing discussions on international cooperation, risk management, and control of AI. Consequently, speaking about these anti-securitization themes is dismissed as naive or strategically unsound.[29]This creates the greatest obstacle to addressing the issue of AI from the perspective of global governance concerning humanity's "survival." In a situation where existential risks such as human extinction or subjugation are foreseen, our future depends on humanity's collective learning capacity to devise coping strategies by drawing lessons from past responses to global catastrophic risks.[30]

The global proliferation of AI technology fundamentally exacerbates the dual dilemma of anarchy versus hierarchy that the existing international order has faced. Furthermore, it can lead to unprecedented levels of transcendent power concentration that human civilization as a whole may find difficult to manage, thereby bringing the issue of "constraint" from a republican perspective to the forefront. ASI, in particular, as a technological agent with the material basis and self-improvement potential to simultaneously acquire omnipotence, omniscience, and omnipresence, could become the ultimate form of tyrannical power that republicanism has historically guarded against. In this case, there is no guarantee that existing alignment strategies or safety designs will function effectively. Consequently, the core of ASI control strategy must converge beyond mere technical risk management towards establishing procedural and structural constraints within the framework of republican constitutional design and public safety. It may even be necessary to include extreme options such as a development moratorium (relinquishment) within the scope of policy imagination. In essence, the advent of the AI revolution is opening a historical juncture that, more radically than the dawn of the nuclear age at the beginning of the Cold War, compels us to re-examine the very sustainability of the anarchic international system. This should be understood as a signal that urges the invention of global republican institutions to ensure that this is not "humanity's last invention."■

[1]Daniel W. Drezner, "Technological Change and International Relations,"International Relations 33 (2), 2019, pp. 286–303.

[2]Ben Garfinkel, "The Impact of Artificial Intelligence: A Historical Perspective," inThe Oxford Handbook of AI Governance, edited by Justin B. Bullock et al., Oxford: Oxford University Press, 2024.

[3]Amelia C. Arsenault and Sarah E. Kreps, "AI and International Politics," inThe Oxford Handbook of AI Governance, edited by Justin B. Bullock et al., Oxford: Oxford University Press, 2024.

[4]Daniel H. Deudney,Bounding Power: Republican Security Theory from the Polis to the Global Village, Princeton: Princeton University Press, 2007.

[5]Daniel Deudney and Devanshu Singh, “Bounding Superpowers: The ASI Control Problem, Public Safety, and Republican Constitutionalism,” in Polycentric Federalism and World Orders, edited by Brandon Christensen, Switzerland: Springer Nature, 2025, pp. 465–502.

[6]Paul van Hooft, Lotje Boswinkel, and Tim Sweijs, Shifting Sands of Strategic Stability: Towards a New Arms Control Agenda, The Hague: The Hague Centre for Strategic Studies, 2022.

[7]Seth D. Baum, Robert de Neufville, Anthony M. Barrett, and Gary Ackerman, “Lessons for Artificial Intelligence from Other Global Risks,” in The Global Politics of Artificial Intelligence, edited by Maurizio Tinnirello, Boca Raton, FL: Chapman and Hall/CRC, 2022, pp. 111–115.

[8]George S. Takach, Cold War 2.0: Artificial Intelligence in the New War between China, Russia, and America, New York: Pegasus Books, 2024.

[9]Jeffrey Ding, “The Rise and Fall of Technological Leadership: General-Purpose Technology Diffusion and Economic Power Transitions,” International Studies Quarterly 68 (2), 2024.

[10]Lee Jae-jun, “US-China Military Technology Competition and Power Transition: Focusing on Artificial Intelligence and Autonomous Weapons Systems,” *Korea and International Politics* 38 (3), 2022.; Office of Science and Technology Policy, “America’s AI Action Plan,” The White House, July 10, 2025, https://www.whitehouse.gov/wp-content/uploads/2025/07/Americas-AI-Action-Plan.pdf.; The White House, “Launching the Genesis Mission,” November 24, 2025. https://www.whitehouse.gov/presidential-actions/2025/11/launching-the-genesis-mission/.

[11]Jeong Gu-yeon, “Military Application of Artificial Intelligence and Strategic Stability,” *Economy and Society* No. 143, 2024.

[12]James Johnson, AI and the Bomb: Nuclear Strategy and Risk in the Digital Age, Oxford: Oxford University Press, 2023.; Vladislav Chernavskikh and Jules Palayer, “Impact of Military Artificial Intelligence on Nuclear Escalation Risk,” SIPRI Insights on Peace and Security No. 2025/06, 2025.

[13]Jacob Stokes, Colin H. Kahl, Andrea Kendall-Taylor, and Nicholas Lokker, Averting AI Armageddon: U.S.-China-Russia Rivalry at the Nexus of Nuclear Weapons and Artificial Intelligence, Washington, DC: Center for a New American Security, 2025.; Jon B. Wolfsthal, Hans Kristensen, and Matt Korda, “Why We Should Worry About Nuclear Weapons Again,” The Washington Post, June 4, 2025, https://www.washingtonpost.com/opinions/interactive/2025/us-russia-nuclear-weapons-proliferation-danger/.; Anton La Guardia, “The Perils of the World’s Third Nuclear Age,” The Economist, November 20, 2024, https://www.economist.com/the-world-ahead/2024/11/20/the-perils-of-the-worlds-third-nuclear-age.

[14]Baele, Stephane J., Iqraa Bukhari, Christopher Whyte, Scott Cuomo, Benjamin Jensen, Kenneth Payne, and Eugenio V. Garcia, “AI IR: Charting International Relations in the Age of Artificial Intelligence,” International Studies Review 26 (2), 2024.

[15]Daniel H. Deudney, Bounding Power: Republican Security Theory from the Polis to the Global Village, Princeton: Princeton University Press, 2007.

[16]Adam Elkus, “Are AI and Democracy Compatible?” Foreign Policy, August 29, 2025, https://foreignpolicy.com/2025/08/29/ai-democracy-dictatorship-agi-governance/ .; George Gilder and Peter Thiel, “COSM Speaker Peter Thiel: The Failures and 'Self-Hatred' of Big Tech,” Mind Matters, October 14, 2021, https://mindmatters.ai/wp-content/uploads/sites/2/2021/10/Mind-Matters-Episode-156-Peter-Thiel-at-COSM-rev1.pdf.

[17]Yuval Harari, *Nexus: A Human History of Information Networks from the Stone Age to the AI Age*, translated by Myung-joo Kim, Seoul: Kimyoungsa, 2024.

[18]Mark Coeckelbergh, “Vulnerability, AI, and Power in a Global Context: From Being‑at‑Risk to Biopolitics in the COVID‑19 Pandemic,” in The Global Politics of Artificial Intelligence, New York: CRC, 2022.

[19]Valentin Weber, “China’s AI-Powered Surveillance State,” Journal of Democracy 36 (4), 2025.

[20]Dean Jackson and Samuel Woolley, “AI’s Real Dangers for Democracy,” Journal of Democracy 36 (4), 2025.

[21]Shoshana Zuboff, *The Age of Surveillance Capitalism: The Fight for a Human Future at the New Frontier of Power*, translated by Bo-young Kim, Seoul: Munhak Sasangsa, 2021.; Mark Coeckelbergh, *Why Artificial Intelligence Is Necessarily Political: The Politics of AI and Freedom, Equality, Justice, Democracy, Power, Animals, and the Environment*, translated by Hyun-seok Bae, Seoul: Saenggak-i-eum, 2023.

[22]Eugenio V. Garcia, “Conclusions: Charting the Challenge of AI IR,” International Studies Review 26 (2), 2024.

[23]Seon-myeong Lee, “I, As Dangerous As Nuclear War/Pandemic… AI Developers Warn of Human Extinction,” https://www.khan.co.kr/article/202305311305001.

[24]Ray Kurzweil, *The Singularity Is Near: When Humans Transcend Biology*, translated by Choong-ho Lee, Seoul: Business Books, 2025.

[25]Nick Bostrom, *Superintelligence: Paths, Dangers, Strategies*, translated by Sung-jin Cho, Seoul: Gakachi, 2017.

[26]Nick Bostrom, “Existential Risks: Analyzing Human Extinction Scenarios and Related Hazards,” Journal of Evolution and Technology 9, 2002.

[27]Nick Bostrom, “The Vulnerable World Hypothesis,” Journal of Evolution and Technology 10 (4), 2019.

[28]Dan Hendrycks and Mantas Mazeika, “X-Risk Analysis for AI Research,” arXiv, September 20, 2022, https://doi.org/10.48550/arXiv.2206.05862.; The deeper fear underlying AI apocalypticism can be captured by Max Weber’s concept of the iron cage. That is, AI can be imagined as the culmination of the runaway or uncontrollable nature of modern instrumental reason. In other words, the question of AI apocalypticism is actually intertwined with the concern that modernity itself is impoverishing our subjectivity, substantive rationality, and reflectiveness. Jay A. Gupta, “Welcome to the Machine: AI, Existential Risk, and the Iron Cage of Modernity,” Telos 203, 2023, pp. 163–69.

[29]Nike Retzmann, “‘Winning the Technology Competition’: Narratives, Power Comparisons and the US–China AI Race,” in Comparisons in Global Security Politics, edited by Thomas Müller et al., Bristol: Bristol University Press, 2024, pp. 237–56.

[30]Seth D. Baum, Robert de Neufville, Anthony M. Barrett, and Gary Ackerman, “Lessons for Artificial Intelligence from Other Global Risks,” in The Global Politics of Artificial Intelligence, edited by Maurizio Tinnirello, Boca Raton, FL: Chapman and Hall/CRC, 2022.

■Author: Tae-seo Cha_Professor of Political Science and International Relations, Sungkyunkwan University.

■ Responsible and Edited by: Jaehyun Lim_EAI Research Fellow

    Inquiries: 02 2277 1683 (ext. 209) | jhim@eai.or.kr