The Silicon Curtain: How Data Colonialism fuels global inequality and conflict

Beneath the Silicon Curtain: A lone figure stands before a looming digital divide, symbolising how data-driven empires cast long shadows over global equity and autonomy.

Introduction – A New Curtain Descends on the World

In 1946, Winston Churchill warned of an "Iron Curtain" descending across Europe — a divide not just of borders, but of ideologies. It marked the beginning of the Cold War: a bipolar world shaped by the tense standoff between the United States, the self-proclaimed leader of the free world, and the totalitarian regime of the Soviet Union. The stakes were existential, the weapons nuclear, and the ideological lines sharply drawn.

Today, many argue we are entering a multipolar world, with the rise of China, India, and the demographic explosion in Africa challenging Western dominance. Scholars like John Mearsheimer have characterised the Cold War as a period of stability born from bipolar equilibrium — but the emerging global order may be far more unstable, fractured, and opaque.

I take a different position — one echoed in Yuval Noah Harari’s Nexus: a new kind of curtain is falling. Not one made of concrete and steel, but of code, algorithms, and firewalled data empires. A Silicon Curtain.

This new divide is being built not with tanks but with training data, proprietary models, and incompatible platforms. As it stands, the United States and China are the dominant AI powers, with Europe and the Global South playing catch-up. In my previous article NarnA.I.: The Stag, The Chicken and The Prisoner’s Dilemma, I explored the game-theoretical logic behind the AI arms race. But this article asks a deeper question:

Will AI lead us to a unified global order — fulfilling the early techno-optimist dream of a decentralised, democratic internet?Or will it shatter the human species into incompatible digital tribes, with algorithmic rule reinforcing the ideological clash between liberal democracy and digital totalitarianism?

My central thesis is this: data colonialism — the invisible, systemic extraction and exploitation of human experience by dominant AI superpowers — is not only reshaping the global order, it is entrenching a new kind of structural inequality. One that will have horrific consequences for democracy, development, and human well-being across the globe.

The Rise of Digital Empires

In the 19th century, the Industrial Revolution gave birth not only to steam engines and railways, but to new global empires. These empires extended their reach across continents by laying down physical infrastructure — railroads, telegraph lines, and steam-powered ships — enabling the movement of people, goods, and military power at unprecedented scale. The logic of imperialism became entwined with the technology of the time, and the extraction of resources and labour from colonies was justified through the machinery of industrial progress.

Today, we are witnessing the rise of a new kind of empire — one built not on steel and coal, but on data and algorithms. Just as railways once defined geopolitical power, digital infrastructure — from cloud servers to AI models — has become the backbone of global influence and economic development. But unlike the empires of the past, today’s imperial builders are not nation-states, but private tech giants: Google, Amazon, Meta, Alibaba, Baidu. These firms have acted like modern-day railway barons, racing to establish control over search engines, online marketplaces, social networks, and cloud infrastructure — the arteries of the digital world.

Where 19th-century empires conquered land, digital empires conquer attention. Their goal is not territory, but total cognitive immersion. They do not require labour camps or plantations — instead, they mine behavioural surplus: the trails of data we leave behind through every like, swipe, purchase, message, or scroll. These fragments of digital life are then fed into machine learning systems that predict, influence, and increasingly manipulate human behaviour — often without our awareness, let alone our consent.

Our participation on the internet is the coal that fuels the AI steam engine.

As Harari notes in Nexus:

“The cat images we uploaded for fun, the selfies we tagged for friends, the reviews we left on products — these were not just cultural noise. They were the raw material for the next empire.”— Yuval Noah Harari, Nexus

One striking example of this transformation was the so-called ImageNet moment. In 2012, a deep learning system known as AlexNet stunned the AI world by achieving unprecedented success in the ImageNet Large Scale Visual Recognition Challenge — an academic competition testing how well computers could learn to identify and classify images. The model was trained on ImageNet, a vast database of annotated images, many of which had been freely uploaded by everyday users across the globe. Harmless, even humourous images — particularly of cats — became the training ground for a revolution.

At the time, it seemed trivial. But this open, crowdsourced data enabled researchers to build neural networks with astonishing accuracy. AlexNet’s breakthrough marked a turning point: machines could now surpass human performance in narrow perceptual tasks. Once that barrier was broken, these architectures were quickly adapted for far more serious applications — including facial recognition, drone surveillance, predictive policing, and social media manipulation.

As Harari writes, “The AI race was on. All the competitors were running on cat images.”

The implications are profound. People around the world unknowingly provided the training data that underpins today’s most powerful AI systems — yet they received no share in the economic value or political power derived from it. Data was harvested globally, processed in American and Chinese data centres, and transformed into commercial and military advantage.

This marks the emergence of what Harari calls a “data-colonial infrastructure” — an invisible, extractive system where digital labour, especially from the Global South, is exploited to enrich the digital elite.

“Raw data will be harvested throughout the world and will flow to the imperial hub... neither the profits nor the power is distributed back.”— Nexus, p. 374

Just as 19th-century imperialism enriched a few nations at the expense of many, digital imperialism today is concentrating wealth, power, and control in the hands of a narrow transnational elite. But this time, the conquest is silent. The borders are invisible. And the empire lives in the cloud.

Data Colonialism: The New Extraction Economy

Whether knowingly or unwittingly, billions of people have immersed themselves into the digital world — a world in which raw data is continuously harvested, often without consent, compensation, or even awareness. Every click, location ping, keystroke, and facial expression becomes a resource: mined, processed, and repurposed by machine learning systems built by distant actors, governed by opaque incentives.

Yuval Noah Harari, in Nexus, describes this process as the emergence of a “data-colonial infrastructure” — a system where individuals and entire nations produce valuable digital labour but remain excluded from its rewards and power. As he puts it:

“Raw data will be harvested throughout the world and will flow to the imperial hub... but neither the profits nor the power are distributed back.”— Nexus, p. 374

This logic mirrors the structure of historical colonialism: the extraction of resources from the periphery to benefit the core. But today, the resources are not cotton, rubber, or gold — they are the patterns of human behaviour, encoded and monetised by algorithms.

The consequences of this new extractive model are felt both within and between nations.

Within countries, as Cathy O’Neil illustrates in Weapons of Math Destruction, algorithms are already reinforcing systemic inequality. Automated systems used for credit scoring, hiring, and policing disproportionately target the poor and marginalised, often without recourse or transparency. In her chapter on education, she highlights how predictive models labelled low-income students as "high risk" — denying them opportunities based on historical patterns, not individual merit. These “weapons of math destruction” are powerful, opaque, and self-reinforcing. As she writes:

“They created their own reality and used it to justify the status quo. That’s what makes them so dangerous.”— Cathy O’Neil, Weapons of Math Destruction

Internationally, the imbalance is even starker. Much of the training data powering today’s AI systems is generated by users in the Global South — where regulations are looser, surveillance is cheaper, and digital infrastructure is often foreign-owned. This data is then processed in high-tech hubs in the U.S. and China, where it is used to train models that drive economic growth, military power, and global influence.

But those who generate the data have no say in how it's used — just as, in the 19th century, rice farmers along the Yangtze River had no influence over the Liverpool–Manchester Railway, even though their exports and labour were vital to the empire it built.

As Harari notes in Nexus, we are living through an uncanny echo of past colonial logic: the separation of production from exploitation; the capture of value in one part of the world, based on invisible labour from another.

And this time, the system is not enforced with muskets or taxes — it is enforced with terms of service, algorithmic opacity, and digital addiction.

Even seemingly harmless participation in social media or image-sharing platforms contributes to the model-building economy. As mentioned, the ImageNet moment — built on publicly sourced images — shows how the infrastructure of AI was crowdsourced globally, often without users realising their cat photos, selfies, and comments would one day feed military-grade surveillance systems or autonomous weapons.

The injustice is not just economic — it is epistemic and political. Those who generate the data are also excluded from the design of the algorithms, the ethics of the systems, and the benefits of their deployment. In this way, data colonialism doesn’t just extract — it silences.

What results is a system where power is hoarded, visibility is one-directional, and inequality is no longer just a product of wealth gaps — it is coded into the infrastructure of digital life.

Digital Dependence and the Illusion of Sovereignty

The word “sovereignty” implies independence — the ability of a nation to govern itself without external interference. But in the digital age, that independence is increasingly hollow. Countries may print their own currencies, run elections, and fly flags at the UN, but beneath the surface, many rely almost entirely on foreign-controlled digital infrastructure to function.

Take the example of Tuvalu, a tiny Pacific Island nation of just over 11,000 people. In 2021, Tuvalu announced plans to build a “digital nation” — a backup copy of itself — because it faces the threat of complete submersion due to rising sea levels. The project was ambitious: digitise land records, cultural artefacts, even governance structures. But it required dependence on cloud infrastructure, data storage, and technical expertise from firms based in Australia, New Zealand, and the United States. Tuvalu may be politically sovereign — but in its most existential project, it is digitally subordinate.

Consider the following thought experiment: If a foreign entity controls your country’s cloud services, internet traffic, or software protocols, do you still possess true sovereignty? If they can see the emails, biometrics, and search history of your political leaders, how independent can your decision-making really be?

The illusion of digital independence is even more fragile in the Global South. Many developing countries outsource their data storage and digital operations to multinationals because local infrastructure is underdeveloped or prohibitively expensive. Surveillance tools are often imported. Payment systems run through Western apps. Even health data — as seen in the aftermath of COVID-19 — was routed through Google and Apple APIs.

But this dependence is not limited to the periphery. Even nations like the United Kingdom — nominally a technological powerhouse — are deeply entangled in foreign digital infrastructure. The UK government hosts many of its services on Amazon Web Services (AWS). Public education tools use Google Classroom. Millions rely on Meta platforms for communication. In 2020, Microsoft beat AWS and Google to win the NHS cloud email and productivity contract, centralising millions of sensitive communications within one U.S.-based provider.

In short, most modern governments cannot function without the corporate infrastructure of foreign-owned tech giants.

If the British Empire once ruled through railways, telegraphs, and gunboats, the empires of today rule through cloud contracts, app ecosystems, and neural networks.

In my earlier article, East India Co. Redux: The Dual Currents of Power and the Endangered Nation-State, I explored this duality in greater depth. On one hand, technology is democratising: open-source models, cheap hardware, and decentralised platforms empower individuals and small actors. But on the other hand, it is deeply centralising: cloud dominance, regulatory influence, and economies of scale are pulling power toward the corporate core.

“Amazon Web Services reinvests billions annually into data centres that fortify its market dominance; Google processes over 90 percent of global searches; and Meta’s social networks touch more than half the world’s population. These firms not only capture value but also set the rules — defining standards, APIs, and even the ethics of data use — so that power flows inexorably toward a corporate center.”

If this framing sounds sensationalist, consider the historical precedent: the East India Company began as a trading enterprise. By the 18th century, it governed vast territories in India, maintained its own standing army, and extracted immense wealth under a corporate flag. In 1803, it fielded a private army of over 260,000 — more than twice the size of the British Crown’s military.

Today’s digital behemoths don’t command battalions, but they do possess regulatory power without representation, jurisdiction without borders, and control without consent.

The East India Company captured land and trade routes; today’s tech giants capture data flows, user attention, and system dependencies.

And just as 18th-century empires claimed they brought “order” and “civilisation” to their colonies, today’s tech firms speak the language of “innovation,” “efficiency,” and “user empowerment.” But behind this rhetoric lies a simple truth: digital dependence is the new form of political subordination.

As the next sections will argue, this growing concentration of digital power is not only reshaping the balance of global influence — it is also rendering traditional notions of national sovereignty dangerously obsolete.

The Silicon Curtain: Fragmentation and Hostile Digital Camps

A new geopolitical fracture is emerging — one that is not defined by traditional borders or even by ideology in the narrow sense, but by data infrastructures, regulatory regimes, and increasingly incompatible digital ecosystems. This fracture represents the gradual descent of a Silicon Curtain across the globe. It is an invisible yet formidable divide that separates nations not through walls or wire, but through algorithms, firewalls, and sovereign code.

In AI Superpowers, Kai-Fu Lee outlines the deepening rift between the United States and China — the two dominant players in the artificial intelligence arms race. While both nations compete for dominance in data, machine learning, and global influence, their underlying systems are diverging in both structure and purpose. In the United States, the AI ecosystem is largely driven by private innovation, venture capital, and consumer behaviour. In China, it is shaped by state-led mobilisation, centralised data control, and a tight integration of commercial and political goals. As a result, each country is erecting a distinct digital empire — not just in terms of values, but in the very architecture of its digital infrastructure.

This divide is already being materially enforced. The United States has banned Chinese firms like Huawei and restricted access to TikTok, citing national security concerns. In return, China maintains its Great Firewall, which blocks American platforms such as Google, YouTube, Twitter, and Facebook. India has banned over 200 Chinese applications, including WeChat and Alibaba’s UC Browser, in response to escalating tensions. In the European Union, concerns over digital sovereignty have led to the formulation of strict data privacy laws like the GDPR and a push for digital infrastructure that is independent of both American and Chinese influence. These are not isolated actions — they are early indicators of a much broader trend: the bifurcation of the global internet into rival and incompatible zones of control.

What we are witnessing is the cocooning of populations into increasingly separate digital realities. This is not simply about echo chambers or personal biases. It is about entire nations — and soon, entire civilisations — consuming knowledge, communicating, and forming worldviews through algorithmic systems that are tuned to fundamentally different values. In China, for example, large language models such as DeepSeek have been trained to avoid criticism of the Chinese Communist Party, to censor references to politically sensitive topics like Tiananmen Square, and to reinforce state-sanctioned narratives. In the West, while platforms may operate under democratic norms, they are driven by corporate incentives that prioritise engagement over truth, polarisation over nuance, and profitability over public good. The result is the same: a distorted reality, governed not by consensus or common knowledge, but by the opaque preferences of the systems in which people are embedded.

Regulation, censorship, and surveillance differ dramatically across borders. In China, the state openly uses AI for social credit scoring and predictive policing. In the West, surveillance capitalism has taken a more insidious form, where consumer behaviour is monetised and manipulated by corporations with little oversight or accountability. Each digital empire trains its AI on different datasets, imbibes different ethical constraints, and serves different masters. The more these systems evolve independently, the less intelligible they become to one another. And so, the dream of a connected global internet — one that could foster a shared human narrative — is quietly dying.

The consequences of this divergence are not abstract. They shape how people understand their own lives, their governments, and their place in the world. A Qatari citizen operating entirely within a Chinese digital ecosystem — using Huawei hardware, Baidu search, and WeChat for communication — may soon have more in common with a Russian citizen using similar tools than with an American using Google, ChatGPT, and Instagram. Not because of culture or geography, but because the algorithmic substrate of their realities is increasingly aligned. Identity is becoming a function not of nation or language, but of digital environment.

This collapse of a shared informational commons has profound implications. When populations are trained on different facts, different histories, and different moral frameworks, cooperation becomes difficult and misunderstanding inevitable. The internet, once hailed as a tool for democratisation and global unity, is turning into a collection of walled gardens, each governed by its own rules and blind to the others. In this fragmented world, the risk is not only mutual surveillance or cyber conflict — it is mutual incomprehension on a planetary scale.

The Silicon Curtain has already begun to descend. And if left unchecked, it may not just divide states — it may divide reality itself.

AI as a Planetary Risk: Inequality, Conflict, and Collapse

It may sound like hyperbole. It may even look like hyperbole. But the truth is, the emergence of the Silicon Curtain and the unchecked spread of data colonialism do not merely pose problems of exploitation or inequality — they carry with them a deeper, more terrifying risk: that even if artificial intelligence remains technically controllable, the fragmentation of human governance could lead to disaster, conflict, and civilisational rupture.

Data colonialism is undoubtedly exploitative. It entrenches structural inequality and incentivises a logic of imperial domination. But these dangers pale in comparison to the more systemic threat — that the absence of global cooperation, the rise of hostile digital blocs, and the reckless pace of AI development will combine into a perfect storm of destabilisation. Artificial intelligence, by its nature, transposes borders. But the institutions tasked with governing it remain trapped within them.

In my earlier article From Safety to Speed: Why Global AI Competition Still Demands Responsible Governance, I outlined how the cautious optimism of the UK’s Bletchley Park AI Safety Summit has given way to an era of geopolitical realism. What began as a multilateral conversation about existential risk has rapidly devolved into an arms race, where speed is valorised and regulation dismissed as weakness. One need only recall U.S. Senator J.D. Vance’s speech, in which he heralded AI as a generational opportunity for American supremacy, while downplaying the need for substantive guardrails. The language of cooperation is being replaced by the rhetoric of national advantage — and the consequences are deeply alarming.

The potential harms are no longer hypothetical. AI-assisted biological research has already demonstrated the capacity to accelerate pathogen design, raising the possibility of synthetic bioweapons emerging not from state labs, but from decentralised actors with the right tools and incentives. Meanwhile, disinformation powered by generative models is flooding our political systems. Deepfakes, automated propaganda, and micro-targeted misinformation are not only distorting public discourse — they are undermining trust, legitimacy, and democratic stability itself. We are entering a world in which truth is no longer an agreed-upon foundation, but a personalised hallucination served by whichever model best confirms one’s prior beliefs.

At the same time, the economic disruptions being ushered in by AI are deep and uneven. Industries essential to national development — from education to finance, logistics to medicine — are being restructured at speed, often with monopolistic concentration in the hands of a few firms and a few countries. Those who lack the compute, capital, or access to cutting-edge models will find themselves permanently locked out of the economic engine of the 21st century. Inequality will not merely widen — it will harden into a global caste system of digital haves and have-nots.

And yet, even in nations that attempt to act responsibly, there is no insulation from global risk. The European Union’s AI Act, despite being hailed as a milestone in regulatory innovation, has no jurisdiction outside its borders. If authoritarian regimes or rogue states choose to ignore best practices, and instead pursue weaponisation or reckless acceleration, the entire world will feel the fallout. This is the logic of climate change, of terrorism, of pandemics: no matter how careful one nation may be, a system-wide collapse requires only a few failure points.

Yuval Noah Harari captures this perfectly in Nexus when he writes, “A world of rival empires separated by an opaque Silicon Curtain would also be incapable of regulating the explosive power of AI.” And opacity is exactly what we are constructing. Proprietary black-box models trained behind closed doors. Open-source systems with no oversight. Press releases dressed up as safety protocols. Military contracts signed in secret. There is no meaningful global AI governance regime — no enforceable treaty, no binding ethical framework, not even a credible mechanism for transparency between superpowers.

And perhaps most dangerous of all, there is no trust. No shared epistemology. No common ground from which to build it.

What we are sleepwalking into is not just a future of uncontrolled superintelligence. It is a world where narrow, short-term national competition renders us collectively blind to the systems we are destabilising. The collapse may come not from rogue machines, but from fractured humans — racing to outdo one another, unable to coordinate, and unaware they are all accelerating towards the same cliff.

This is the planetary risk we now face. Not just inequality. Not just surveillance. But civilisational failure — born not of malevolence, but of a refusal to slow down.

Conclusion: A Call for Digital Decolonisation

Let us hope it will not take another global conflagration — another war in which the individual is reduced to an expendable pawn — for us to recognise that the age of empire is over. Let us pray that it will not require a nuclear catastrophe for us to understand that this digital arms race, cloaked in the language of innovation and progress, is driving us beyond the edge of reason.

The defining question of our time is this: Can we build a digital future that connects, rather than colonises? Can we construct an information infrastructure that empowers the many, rather than enriching and entrenching the few? Or will we allow ourselves to become the raw material for a new kind of empire — one ruled not by kings and flags, but by code, platforms, and opaque algorithms?

If the 20th century was shaped by those who controlled oil, the 21st will be defined by those who control, interpret, and disseminate data. And the deeper question that follows is one of consent: Do we consent to be governed by unelected corporations and their complicit nation-state partners? Do we accept a world where liberty is traded for convenience, truth for engagement, and sovereignty for dependence?

As the United States and China manoeuvre in this high-stakes game of digital geopolitics, both risk backing themselves into a corner — while the rest of us, the so-called expendable pieces, are left drifting in a fog of engineered confusion and algorithmic control.

The call, then, is not just for caution. It is for digital decolonisation — a conscious effort to reclaim autonomy, rebuild trust, and reimagine a global internet that is governed not by empire, but by shared principles, democratic accountability, and universal human dignity.

Because if we fail to act, the future will not be written by people — it will be programmed by powers we never chose, in languages we do not speak, and in interests that are not our own.