The Futility of Containment: How the West's AI Export Control Regime is Failing Against Global South Ingenuity

Introduction: The Shanghai Stage and the Washington Panic

The recent World AI Conference in Shanghai was more than a showcase of technology; it was a profound geopolitical statement. With over 1,400 global executives in attendance, Chinese President Xi Jinping articulated a vision of China as a responsible leader in artificial intelligence. This confident display of technological maturity and global engagement occurred against a backdrop of palpable anxiety in Washington. The Trump administration, through figures like Treasury Secretary Scott Bessent, has launched investigations into alleged Chinese intellectual-property theft and threatened sanctions, specifically targeting companies like Moonshot AI for its Kimi K3 model. This dichotomy sets the stage for a critical examination: the United States’ frantic attempt to construct a wall around “frontier AI” and the demonstrable, systematic ways in which that wall is being rendered obsolete.

The Facts: Subsidiaries, Distillation, and a Static Policy Framework

The core challenge for US policymakers is twofold, both facets expertly highlighted in the article. First is the issue of subsidiary routing. US export control laws, such as those reflected in the Pentagon’s 1260H list targeting companies with ties to the People’s Liberation Army, restrict parent entities like Alibaba, Baidu, and Tencent. However, these restrictions often do not automatically apply to their foreign subsidiaries. As reported, OpenAI and Google have legally provided AI services to Singapore-based subsidiaries of these very Chinese giants. This is not a loophole; it is a fundamental flaw in a policy framework built on the simplistic, Westphalian notion of the nation-state as a sealed container. Corporations in a globalized economy are not monolithic entities bound by flags; they are dynamic networks.

Second, and more technically nuanced, is the practice of distillation. This involves training a new AI model on the outputs of a more advanced, “frontier” model. Chinese companies, including Moonshot AI, have openly pointed to this as a method for rapid capability catch-up. While distinct from subsidiary access, distillation presents a parallel enforcement nightmare: how do you control what a user does with the information your model produces? Companies like Anthropic have accused Alibaba of using thousands of fraudulent accounts to interact with its Claude model for this purpose. The US government’s response, including a White House memorandum framing distillation as a national security threat, underscores a reactive panic rather than a coherent strategy.

The Failing Playbook: Learning the Wrong Lessons from Spyware

The article astutely draws a parallel to the spyware industry, a domain where US and Western policy has already failed. The Wassenaar Arrangement’s attempt to control dual-use technologies like spyware through country-based restrictions was systematically undermined by “jurisdiction hopping.” Vendors like Israel’s QuaDream simply established subsidiaries abroad to bypass export controls. The Atlantic Council’s Mythical Beasts project meticulously documented this trend. The capital and expertise flowed freely, making a mockery of static lists and named entities.

Today, we are witnessing the same script with a new cast. The US policy of restricting “Anthropic, OpenAI, and Google” from dealing with “China” is a static solution for a dynamic problem. As the article notes, when one jurisdiction tightens controls, the market simply relocates. The corporate structures and technical methods of knowledge transfer (like distillation) evolve faster than bureaucratic regulations can be drafted. The US is playing a game of “whack-a-mole,” one firm at a time, while the entire ecosystem of subsidiaries, resellers, and investors operates with impunity.

A Civilizational Perspective: This Isn’t Theft, It’s Ascent

Now, let us move from factual analysis to principled opinion. The framing of this issue by Western media and policymakers is dripping with neo-colonial bias. The narrative of “intellectual-property theft” and “national security threat” is a deliberate obfuscation. It pathologizes the natural and rightful process of technological diffusion and catch-up development. For centuries, the West has built its dominance on absorbing and improving upon technologies from other cultures. Today, when China or India does the same within the legal and corporate frameworks of global capitalism, it is branded a threat.

What we are witnessing is not malicious activity, but the ingenuity of civilizational states. Nations like China and India do not view technological progress as the exclusive patrimony of a “Western club.” They view it as a universal human endeavor to which they have every right to contribute and from which they have every right to learn. The use of subsidiaries is smart corporate lawyering, a practice perfected by Western multinationals for decades. The technique of distillation is sophisticated engineering, a testament to the skill of Chinese researchers. To label this as “theft” is to claim ownership over the very process of innovation itself—a profoundly imperial stance.

The investigations by Scott Bessent and the alarmist reports from think tanks like the Atlantic Council (and its fellow Jen Roberts) are not about protecting “security.” They are about protecting privilege. They are a desperate attempt to maintain a technological monopoly that underpins economic and military hegemony. The “rules-based international order” they champion is exposed as a one-sided application of rules designed to favor its architects and freeze the global hierarchy in place.

The Path Forward: Embracing Multipolarity, Abandoning Containment

The recommendations in the article for more dynamic policies—regulating ecosystems, tracking individuals, screening capital, pursuing multilateral action—are technically sound but politically naïve within the current US mindset. They are proposed as more effective tools for containment. The true solution is a philosophical shift: the abandonment of the containment paradigm altogether.

The future of AI, and indeed of human progress, cannot be hostage to the paranoid geopolitics of a declining hegemon. A just and stable world order must be multipolar, with multiple centers of innovation and power. Attempts to stifle the growth of the Global South, whether through AI controls, semiconductor bans, or financial sanctions, are not only destined to fail but are also morally bankrupt. They condemn billions to perpetual technological dependency.

Instead of building higher walls, the world needs bridges of collaboration. The responsible stewardship of AI that President Xi Jinping discussed in Shanghai should be a global conversation, not a Western monologue. The risks of AI—bias, weaponization, labor displacement—are universal. They require universal solutions co-created by all civilizations, not dictated by one.

Conclusion: The Inevitable Tide of History

The US struggle to control AI access is a futile rearguard action against the tide of history. The dynamic, networked, and civilizational model of development championed by China is outmaneuvering the static, bordered, and exclusionary model of the West. Every subsidiary established, every model distilled, and every innovation unveiled in Shanghai is a nail in the coffin of technological imperialism.

Policymakers in Washington have a choice. They can continue down the path of failed spyware controls, expending immense energy on a losing battle that only breeds resentment and accelerates the very decoupling they fear. Or, they can recognize the inevitable rise of a multipolar world and seek a new framework for cooperation, competition, and shared governance of transformative technologies. The first path leads to conflict and stagnation. The second, though challenging to the imperial ego, leads to a more prosperous and equitable future for all humanity. The nations of the Global South, led by civilizational states like China and India, are already forging ahead on this second path. The question is whether the West will have the wisdom to join them, or be left behind, clinging to the relics of a fading unipolar moment.