Boris Babic

Brian Wong Yue Shun

Reviewed by Dr. Eman Abdel HalimDespite growing interest in the geopolitical implications of artificial intelligence, the subject remains relatively unexplored in the academic literature to date. Geopolitics of Artificial Intelligence, published in September 2026 as part of the Cambridge Elements in Philosophy and AI series, seeks to address this gap. The book is authored by Boris Babic and Brian Wong Yue Shun, professors and researchers at the University of Hong Kong whose work focuses on the philosophy of science and artificial intelligence, as well as the ethical and legal dimensions of algorithmic systems and their various applications.Drawing on this academic background, the authors argue that artificial intelligence is no longer merely a tool employed by states in traditional geopolitical competition. Rather, it has become a force reshaping the nature of global power, the ways in which it is exercised, and the mechanisms through which it is legitimized. Against this broader understanding, the book provides a philosophical and analytical examination of the growing intersection between AI and geopolitics, with particular emphasis on the ethical and legal implications of using this technology in international relations. The authors also propose a framework centered on greater international justice and strategic cooperation, rather than reducing competition over AI to a "zero-sum arms race" among major powers.The Ethical DimensionThe authors outline four main reasons why the ethics of artificial intelligence warrant greater attention in relations between states:1. Balancing Benefits and Harms: The rapid proliferation of AI technologies has created an unprecedented range of potential benefits and harms. In some respects, these resemble the effects of earlier technologies that transformed balances of power and relations between states, such as the atomic bomb and the internet, while differing from them in other ways. This makes it necessary to examine the normative implications of these benefits and harms,&nbsp;or of the risks and opportunities arising from the use of AI technologies.According to the authors, this requires confronting a world characterized by scarce resources and competing claims and interests, making the equitable distribution of AI technologies particularly important. There is therefore a need to establish normative principles for balancing competing risks and opportunities, providing an important rationale for developing a theoretical framework for AI ethics in a geopolitical context.2. The Global Justice Dilemma: The rise of AI highlights many of the tensions between competing values that feature prominently in contemporary philosophical literature on global justice. Most government policies, guidelines, and platforms related to AI governance, for example, tend to focus heavily on protecting states’ core interests, potentially prompting governments to impose tighter restrictions on AI technology exports. The paradox, however, is that some restrictions advocated by Western liberal democracies could, in certain cases, undermine the interests of their own citizens as well as those of citizens in allied countries.At the international level, the authors argue that smaller, developing, and poorer economies have an interest in ensuring that technologically advanced economies do not restrict exports of essential components needed for AI development. Technologically advanced states, however, may resist efforts to secure greater access on national security grounds, fearing that advanced technical knowledge could be weaponized against them by actors that are considered a security risk. This reasoning has repeatedly been used to justify restrictions imposed by China, the United States, and the European Union on exports of certain components in their respective AI supply chains. These tensions underscore the need to address the global challenges posed by AI in interstate relations while protecting the sovereignty and interests of fragile and developing states that may face restrictions on access to and use of AI.3. The Challenge of Responsibility and Accountability: The use of AI technologies in warfare, in particular, complicates the attribution of responsibility for decisions and actions involving these systems. In this context, the authors point to the Israeli military’s alleged use of the AI-based identification system "Lavender" during the war in the Gaza Strip, which was reportedly used at one stage to identify 37,000 potential targets based on suspected links to Hamas. They also cite India’s use of AI-enabled drones during its 2025 border conflict with Pakistan.Regardless of the authors’ assessment of the ethical justification for such uses in warfare, these examples raise clear questions about responsibility for developing and deploying AI technologies in military operations. The authors argue that responsibility extends beyond the states and militaries using these technologies to include the engineers, investors, and companies that develop or finance them, particularly when decisions are made about whether advanced technologies should be made available or sold to militaries and intelligence agencies. This broader conception of responsibility has implications for the potential military uses of AI, including whether restrictions should be placed on its development or availability to military and security actors.4. Gaps in Existing Theoretical Frameworks: The authors argue that conventional frameworks to international relations and geopolitics pay insufficient attention to the ethical questions raised by artificial intelligence. At the same time, the existing literature on AI ethics does not fully capture the tunique geopolitical challenges posed by the technology, which include the risks of conflict escalation, unjustified violence, and the misallocation of resources. This highlights the need for an integrated approach that introduces ethical considerations into the geopolitical analysis of AI.Rejecting the "Zero-Sum" ApproachThe authors offer a philosophical and structural critique of the "zero-sum game" that, in their view, shapes how some major powers approach competition over artificial intelligence. They argue that treating technological development as a contest in which one side can gain only at another’s expense overlooks the opportunities for cooperation and mutual benefit that AI could provide. Their main arguments can be outlined as follows:1. Cooperation as a Rational Interest: The book devotes a mathematically oriented chapter to demonstrating that cooperation and narrowing technological gaps can be rationally beneficial for states, challenging political narratives that portray AI competition as a zero-sum race similar to the space race, in which success is defined by outperforming the other side. The authors argue that this model does not fully apply to artificial intelligence because technological advantage is not established at a single decisive moment but evolves continuously as capabilities advance.The book further argues that cooperation, even when unilateral, may prove more beneficial than mutual isolation. The authors cite the example of Chinese company DeepSeek, which made its R1 model open source, helping expand its global adoption and strengthen the company’s presence in the U.S. market. Conversely, they argue that stringent export restrictions on certain AI technologies, such as NVIDIA processors and chipmaking equipment, may deprive companies of economic returns while encouraging competitors to accelerate the development of independent supply chains. According to this argument, such restrictions could ultimately weaken the exporting country’s influence and its ability to shape global regulatory standards.2. Gains Are Not Necessarily Zero-Sum: The book rejects the assumption that progress by one country in artificial intelligence necessarily entails losses for another. From a rational-choice perspective, the authors argue that greater cooperation and more balanced access to technology can produce mutual benefits. In this context, they emphasize the importance of consistent AI regulation, the safe and transparent development of its applications, and broader access to the technology, which could enable groups lacking sufficient economic resources and capabilities to benefit from it.By contrast, the authors argue that intense interstate competition could produce mutually negative outcomes in what is known as a non-cooperative "Nash equilibrium." Such an outcome could cause states to miss out on the potential gains from trade and the exchange of technical knowledge that otherwise would have been realized, while also limiting their chance to provide public goods that depend on international cooperation.3. Expanding Public Goods: The authors challenge the view that cooperation is inherently risky, instead drawing on Jean-Jacques Rousseau’s "stag hunt" model, which suggests that cooperation among actors can generate greater benefits than each could achieve independently. Applied to artificial intelligence, this logic suggests that interstate cooperation could help develop solutions for global challenges such as climate change and pandemics. By contrast, a narrow focus on individual national interests could prevent states from achieving these shared benefits.The book therefore suggests that states may face two broad choices: either share AI resources and technologies or exclude other actors from them. Reciprocal sharing has the potential to produce mutual gains through trade and cooperation on governance, whereas reciprocal exclusion limits the potential benefits of technological exchange.4. The Growing Role of Middle Powers: The authors caution against an excessive focus on the role of major powers that underestimates the importance of small and middle powers, as well as non-state actors, in shaping the global AI landscape.They argue that these actors can play an important role by building multilateral partnerships and avoiding dependence on a single sphere of technological influence. In this context, the authors cite Singapore and Vietnam as countries that pursue hedging strategies aimed at avoiding full alignment with any major power in developing their AI capabilities. They also highlight Malaysia’s experience in strengthening its position in semiconductor manufacturing and re-exports by leveraging its ties with both U.S. and Chinese companies operating in the sector. The authors argue that this approach gives Southeast Asian countries greater autonomy and flexibility in navigating technological competition among major powers.5. The Risks of Technological Isolation: The authors argue that efforts to develop closed forms of "sovereign AI," driven by a zero-sum mentality, could result in more isolated systems that are less able to benefit from the global diversity of data and knowledge. They also suggest that restricting the flow of data among different actors could lead to technical problems such as "concept drift," as well as potentially increasing certain forms of bias in large language models (LLMs).According to the authors, these problems could affect the quality of political and military decisions based on AI systems, potentially increasing the risks of miscalculation and unintended conflict escalation. The risks arising from technological isolation may therefore extend beyond any individual state and affect the broader international environment.Potential SolutionsHaving made the case for greater international cooperation on artificial intelligence and reducing barriers to access to advanced technologies, the authors propose several recommendations, most notably the following:1. Establishing International Dispute-Resolution Mechanisms: Given the importance of developing standards and institutions that promote interstate cooperation in the use of AI technologies, build trust, and reduce misperceptions among international actors, the authors call for the establishment of a specialized mediation and arbitration organization or center to resolve geopolitical disputes related to AI supply chains, including those involving semiconductors, rare earth minerals, and algorithms. They suggest that such an institution should be located in a geographically and politically environment, similar to proposals for the International Organization for Mediation (IOMed).The authors argue that dispute-resolution mechanisms can strengthen cooperation by fostering the trust necessary for international actors to believe that their counterparts will behave cooperatively—an important prerequisite for expanding international collaboration. In other words, effective, accessible, and low-cost dispute-resolution mechanisms could improve the prospects for cooperation by ensuring that non-cooperative behavior can be addressed quickly.2. Reforming Supply Chains and Strengthening Multilateralism: The authors argue that some of the tensions between countries over the AI build-out stem from supply constraints and unequal access to technology. From their perspective, ensuring access to rare earth minerals, semiconductors, chips, and graphics processing units (GPUs) is therefore an important element of strengthening international cooperation.In this context, they call for reducing reliance on unilateral restrictions and export bans, such as U.S. export controls on graphics processors and lithography equipment, in favor of multilateral supply agreements that ensure the flow of resources and reduce incentives for states to pursue forced self-sufficiency. They also emphasize the importance of enabling small and middle powers, as influential actors in a multipolar international system, to maintain a degree of technological autonomy and flexibility and avoid being forced into full alignment with a single bloc. Such an approach could help prevent the global technological system from fragmenting into competing blocs—or what the authors describe as a "digital iron curtain."3. Strengthening Accountability and Limiting the Concentration of Power: In the absence of a comprehensive international legal authority with uniform powers to impose and enforce sanctions for AI-related risks, the authors propose developing national policies and legislation alongside stronger international coordination. Such measures could boost accountability while helping avert a scenario where geopolitical power and decision-making becomes concentrated among a small number of major technology companies.This approach is accompanied by a call for "forward-looking responsibility frameworks" designed to anticipate potential risks and ethical, social, and technological consequences before they materialize, allowing developers and operators to address potential harms arising from their systems at an early stage. The authors also propose developing accountability and compensation mechanisms for cases in which AI systems cause cross-border harm, including algorithmic audits, financial compensation, and other measures. Rather than limiting responses to such harms to traditional criminal proceedings, these mechanisms could help accelerate fact-finding and strengthen protections for those affected.Overall, Geopolitics of Artificial Intelligence offers an approach that goes beyond viewing AI simply as an arena for technological and military competition among major powers. Instead, it examines the technology’s philosophical and normative dimensions and its implications for political philosophy, rational choice theory, and the ethics of international relations. Building on these perspectives, the book emphasizes the importance of developing institutional and legal frameworks that can strengthen trust and multilateral cooperation while mitigating the risks of technological competition. Such frameworks could help manage the potential risks of artificial intelligence while maximizing the benefits it can provide at the international level.

How Can the Benefits of Cooperation Be Maximized in the Age of Artificial Intelligence?