Brokerage in the Black Box: Swing States, Strategic Ambiguity, and the Global Politics of AI Governance

In the twenty-first century, artificial intelligence has moved out of science fiction and into the center of global power. It is no longer just a tool for writing emails or recognizing faces; it is a strategic asset that nations are racing to control. The world is currently divided into two main camps: one led by the United States, which favors a market-driven approach where private companies lead the way, and another led by China, where the government directs development for national security and industrial goals. Between these two giants lies a complex web of trade, technology, and finance that makes it impossible for most countries to simply pick a side and walk away. This creates a difficult situation for smaller, technologically advanced nations. They must navigate a landscape where the technology itself is often a mystery, even to the people who build it. These systems, known as "black boxes," make decisions based on patterns so complex that their inner logic cannot be fully explained. This lack of clarity is not just a technical glitch; it is a permanent feature of modern AI, driven by the sheer scale of the data used and the desire of companies and governments to keep their methods secret for competitive or security reasons.

A new study by Ha-Chi Tran at the London School of Economics and Political Science explores how certain countries are turning this confusion into power. The research focuses on a group of nations the author calls "technological swing states." These are countries that are not superpowers but possess enough technical skill and strategic flexibility to act as middlemen. The paper examines three specific examples: South Korea, Singapore, and India. The central finding is that these nations are not just passive observers waiting for the US and China to fight it out. Instead, they are actively using the uncertainty and opacity of AI to their advantage. By refusing to make permanent, irreversible commitments to one side, they create space to negotiate, build trust, and shape the rules of the game. The study suggests that because no one can fully see inside the AI systems of their rivals, these middle powers can step in to provide the trust and verification that the world needs, effectively becoming the brokers of the new digital order.

The research begins by redefining the problem of AI opacity. For a long time, experts hoped that we could simply make these systems transparent, forcing them to explain their reasoning in plain language. However, the paper argues that this is likely impossible for the most advanced systems. As AI models grow larger, with hundreds of billions of connections, they develop behaviors that cannot be traced back to simple causes. Even the engineers who build them cannot always explain why a specific decision was made. Furthermore, there is a strong economic and political incentive to keep these systems secret. Companies want to protect their trade secrets to stay ahead of competitors, and governments want to hide their capabilities for national security. The paper suggests that trying to force total transparency is a losing battle. Instead, the focus has shifted to "institutional transparency." This means that rather than trying to understand the black box itself, we rely on checking the process around it. We look at who is auditing the system, what rules they followed, and whether there are independent bodies to certify that the system is safe. This shift turns the inability to see inside the machine into a new kind of political leverage.

South Korea illustrates the first strategy used by these swing states: delay and hedging. As a major producer of the computer chips that power AI, South Korea sits right in the middle of the supply chain between the US and China. It cannot afford to fully align with the US without angering China, its largest customer, nor can it fully side with China without losing access to American technology and security. The paper shows that South Korea has learned to use this position to its benefit. When the US introduced strict rules on selling chip-making equipment to China, South Korea did not immediately choose a side. Instead, it negotiated temporary exemptions and delays. This strategy of waiting and keeping options open allows them to avoid the high cost of picking a permanent winner in a race that is still being run. By staying ambiguous, they force both the US and China to treat them with care, knowing that pushing them too hard could cause them to drift toward the other side. This hesitation is not indecision; it is a calculated move to preserve their freedom of action in a volatile environment.

Singapore takes a different approach, which the paper calls "selective alignment." Unlike South Korea, which is deeply tied to heavy manufacturing, Singapore's power comes from its role as a hub for data and digital services. It acts as a trusted connector in the region, hosting data centers and setting standards for how AI should be governed. Singapore does not wait for the US and China to agree; instead, it builds its own frameworks for safety and ethics that both sides can accept. It creates rules for how AI systems are tested and audited, offering a form of verification that the world can trust even if it cannot see the code inside the machine. By doing this, Singapore becomes essential. Countries and companies that want to operate in Southeast Asia must follow Singapore's rules, and global powers need Singapore's approval to show they are acting responsibly. This allows a small nation to punch above its weight, shaping the global rules of the road without ever having to choose a single superpower as its master.

India represents a third strategy known as "normative intermediation." While Singapore focuses on the rules of the game, India focuses on the meaning of the game itself. India is a massive player in the software world and has deep ties to both the developing nations of the Global South and the advanced technology alliances of the West. The paper argues that India uses this position to change the conversation. Instead of just arguing about safety or security, India pushes for a narrative of "AI for All" or "AI for Development." By framing the discussion around helping poorer nations grow, India creates a new set of goals that the superpowers must address to remain legitimate. India does not just translate existing rules; it rewrites the questions being asked. When the world debates what AI should do, India's voice helps define the answer, ensuring that the needs of the developing world are not ignored. This gives India a unique kind of influence, allowing it to bridge the gap between different worlds and set the agenda for how AI is used globally.

The study concludes that these three nations are not merely reacting to the rivalry between the US and China; they are actively shaping the outcome. By exploiting the fact that AI is opaque and that the world is divided, they have found ways to turn uncertainty into power. They use delay, selective partnerships, and new ideas to keep their options open and to ensure that they are not left behind. The paper suggests that this is not a temporary situation but a structural feature of the modern world. As long as AI remains a black box and the great powers remain in competition, these middle powers will continue to act as the essential brokers, deciding who gets to play, how the game is played, and what the rules actually mean. The research does not claim that this is a perfect solution or that these countries are immune to pressure, but it does show that in a world of hidden algorithms and competing empires, the ability to navigate the gray areas is the most valuable asset of all.