Britain deceives itself. We deceive ourselves about our history, and we deceive ourselves about the lessons of that history. Orwell compared Forties England to a Victorian family, its cupboards bursting with skeletons, its members exhibiting “a deep conspiracy of silence” about the sources of their income; today we talk endlessly about that income and of the immorality of empire, but still we derive the wrong lessons. As we face the era of AI, Britain’s imperial history offers a stark warning to the present: that the result of vassalage is destruction.
Consider the weavers. In 1750, India produced 25% of the world’s textiles. Beloved by consumers, the fine muslin cloths of Bengal were the envy of manufacturers throughout the globe. Braving the vagaries of wind, shipwrecks, piracy, and disease, generations of Dutch, English, and Portuguese ships made the dangerous, year-long journey to fetch this precious cargo. Concentrated in Bengal, Coromandel, and Gujarat was a rural and decentralised workforce of skilled weavers, their expertise passed down from father to son. These weavers were the backbone of India’s export system. Nevertheless, within a few decades they had been reduced to poverty. By 1830 India had become a net importer of textiles from Britain; by 1900, it produced a mere 2% of global output. Its textile trade had collapsed, and with it the livelihood of the weavers. Previously prosperous, textile-producing regions struggled with repeated famines; a world-leading industry had been wiped out.
Three factors reduced India to poverty: loss of sovereignty, the power of a foreign multinational, and the impact of automation. Amid the ruins of Mughal power, the merchants of Britain’s East India Company seized control of key ports and provinces. Supported by the British state, they used their new power to squeeze out French and Dutch trading interests — thus wrecking the economic leverage of the weavers, who could no longer choose their export market or set their own prices. A decade or two after the final conquest of India, British textile firms — protected for years from Indian competition due to a strong tariff regime — began utilising technological innovations such as the spinning jenny, the spinning frame, and Crompton’s mule. Combined with steam power, these undercut, outcompeted, and ultimately destroyed India’s textile industry. The present now echoes the past; we would be wise to listen.
Bengal’s past is Britain’s future. They may not know it, but the lawyers, financial experts, and account managers of the British service economy are almost as exposed as the Indian artisans once were. The things that brought catastrophe to Bengal — the erosion of local sovereignty, the state-backed power of foreign corporations, and the wrenching effect of automation overseas — now threaten the economy and society of the United Kingdom. If we do not act, we will face ruin.
To understand why, reflect on the world around you. To be alive, connected to the internet, and minimally curious is to experience the dizzying advance of technology. On any given day we may use ChatGPT or one of its competitors as an accountant, therapist, copywriter, spellchecker, graphic designer, legal drafter, doctor, business analyst, or artist. Entire industries rise and fall in the sociological blink of an eye; influencers, crypto traders and SEO analysts work in industries that didn’t exist 20 years ago, yet now stand to be automated away. More satellites wheel overhead every month, connecting us at all times to a world of apps, communication, and business; electric vehicles are transforming global supply chains; drones are revolutionising warfare; robots clean, dance, drive, and compete in marathons.
Soon, possibly within a decade, humanity may create AI systems capable of outperforming humans at every task we do. Even should full-on “AGI”-type systems not arise, huge numbers of roles will be automated at a pace and speed unlike anything in history. This will have epochal effects on every part of our society. The optimistic view is that this automation will bring with it a wondrous array of new roles and opportunities for displaced workers; pessimists point to the disruption and turmoil experienced by the workers who live through such changes, the sheer disruptive pace of this automation, and how it will impact social power. Individual human workers will stand in an increasingly unequal relationship to the new mechanisms of production; this will impact every element of society.
Britain’s workers are highly exposed. Our lawyers are finding that AI can be used to generate and review contracts, and our accountants have discovered ChatGPT can be used to analyse accounts. DeepResearch — an AI agent integrated into ChatGPT — does the work of a research team; Gemini can easily do the work of a set of graphic designers. Our recruiters and HR teams are increasingly composed of chatbots. And yet Britain doesn’t own any of this technology.
The fact of automation changes the social and international balance of power. Even the most monstrous tyrannies have ultimately relied upon consent. Had Russia’s people risen up as one, Stalin’s regime would have ended in a day. Without workers, factories stop, mines stand still, and ships go unloaded — but in an automated economy, this leverage will no longer exist. Cognitive and physical automation will mean that the contribution of human workers is less and less important; more and more relative power will thus accrue to the owners of AI-enabled systems. The risk is not that mass resistance will be futile; it may simply become irrelevant.
This startling future is being driven by a small subset of US and Chinese companies. Like the East India Company of two centuries ago, the largest tech companies have a power and reach that surpasses most sovereign states. This power not only threatens the independence of individual states; it threatens the viability of the Westphalian order. The companies’ decision-making shapes the sovereignty of entire nations, and they increasingly control the terms under which businesses make it to market, decide public health, and influence political processes. Indeed, tech firms control the very medium of political action. A tweak in an algorithm of a search platform, social media company, or AI chatbot can lead to the rise and fall of governments, and a decision to block or boost news items and politicians decides what is discussed and who gets to discuss it.
Decisions made by the tech majors can bring entire countries to their knees. Rejecting some $290 million in regulatory fines, Meta is currently threatening to withdraw from Nigeria; such a move would devastate the millions of small- and medium-sized businesses that rely on Facebook and Instagram for their marketing. In many places, national governments are relying upon foreign corporations to achieve their basic functions.
Matters of war and peace are now decided by the whims of tech CEOs. Through Starlink, Elon Musk decides the success or failure of military actions in Ukraine. By providing Ukraine with Starlink access in 2022, he gave it vital communication systems and thus kept the country in the war; by refusing Ukraine access to Starlink over Crimea, he prevented planned offensives upon Russia.
Covid even showed the tech majors dictating terms to nuclear powers. Utilising the market power of their app stores, Apple and Google were able to dictate the Covid responses of Western economies. These companies determined which apps could be offered to citizens, how many could be developed per region, and how they could be used. Britain and France — top-10 economic powers, members of the Security Council, and in possession of nuclear weapons — were forced to bow before the joint determination of Apple and Google to stop them using centralised databases within their national contact tracing apps.
Generative AI further deepens the power of the tech majors. A new generation of tech giants — OpenAI, Perplexity, and others — are aiming to own the future. By controlling the foundational models from which LLM-based applications are being built, they have become fundamental economic actors. In an economy that uses AI agents, OpenAI and its competitors will reap outsized profits from replacing workers. With this, a new, yet deeper era of corporate power is being born.
Like other middle powers, the UK must work out how to tax, regulate, and limit this power, or even build its own — but it does not act upon a level playing field. US policy strives to reshape the world to the benefit of its tech companies. Barriers to profitability are being swept aside, new markets are being opened up, competitors and their products isolated, and tariffs used as a tool of policy.
A crucial factor behind the EIC’s conquest of India was the support it had in Britain’s parliament. Through bribery and lobbying, the EIC knew it could rely on Britain’s military, diplomatic, and financial support. Recall now Trump’s second inauguration day, and the spectacle of the tech CEOs. Bezos, Musk, Zuckerberg and Pichai — the four wise men, or, at any rate, exceedingly rich ones — lined up to do homage before the king. Having rejected and resisted Trump 1.0, they see in Trump 2.0 an opportunity to control the future.
The US government now desires to prevent the implementation of AI safety rules, and to eviscerate the tech taxes of Britain and Europe. Despite the recent US-UK trade deal, the US is still pushing to scrap the UK’s digital services tax. EU taxes are another target; even before Trump’s election, the US used the threat of punitive tariffs to pressure the EU over digital taxes. Nationally, Musk’s DOGE is being given free rein to bring AI into the core functions of the US state; GOP policymakers, meanwhile, are attempting to freeze all state-level control or regulation over AI, a move that would primarily benefit US tech majors.
Barely reported in the UK press, Trump’s recent dealmaking in the Persian Gulf shows that the full weight of US diplomacy is being applied to capture markets for AI. Travelling with Trump on his Gulf tour were CEOs of NVIDIA, Amazon, OpenAI, IBM and Tesla; the results are staggering. Inking some $2.2 trillion in technology deals for US companies, Trump has reversed the “AI diffusion rule”, a Biden-era policy placing export restrictions on AI chips and model weights. Instead, the US now gambles that by flooding Saudi Arabia, the UAE and Qatar with American chips and technology, it will bind them to the US and US companies.
Central to all of this is China. Just as competition with France drove Britain and the East India Company to expand their international reach, so too does US competition with China guide its priorities in diplomacy and technology. The US aims to cut China out of international markets, and will use all diplomatic and commercial means at its disposal. This might help the US, but it might not help its allies; captive markets are more easily controlled.
The Gulf has leverage: vast pools of petro-capital, cheap land, cheap energy, and active strategic competition between China and the USA. Britain does not; our land is scarce, our energy is expensive, and we are driving away capital. As an unreflective vassal of the US, we rarely consider the geopolitical implications of limiting our strategic supply base; whatever the realities of Huawei’s security implications, the scandal showed the UK as content to reject Chinese tech providers on the basis of US pressure. In doing so we are reducing our options. A world where China is cut out of international tech markets is a world where middle powers have less choice and less leverage.
It is fantasy to believe that the gains from AI will be fairly distributed. The surging power of US and Chinese tech multinationals layers itself upon a world of deep inequalities, tensions, and rivalries between and within individual countries. We cannot presume that the shareholders of OpenAI or DeepSeek possess a deep urge to prop up the polities of Europe, or that the immiseration of once-prosperous regions will move the conscience of AI leaders. Universal Basic Income is proposed as a key tool for managing the problem of mass unemployment. But if only a handful of states and companies control the most valuable AI systems, and if they use their market power to reduce their local taxes, other governments will have no way to pay for it. Only by building and scaling homegrown AI companies, and by offering value to the technology industry as a whole, can Britain hold a stake in the future. Britain must build leverage, not loyalty.
Britain has more options than the weavers. To survive, we must do four things: embrace AI, develop independent national capacity, increase our leverage, and articulate our national interest. The clock cannot be turned back; we must embrace a new age of ambition. Driving forward the adoption of AI will boost the growth and productivity of the British economy, and expand the market for homegrown AI companies. Every cultural and social measure should be considered: prizes, competitions, TV shows. Bring back Robot Wars, bring in Robot Marathon, or Robot Bake-offs; every schoolchild should aspire for their robot to compete, and win. Small businesses must be encouraged to embrace AI; towns and villages should compete to experiment, and we should mobilise groups like the Cambridge Ring as providers of insights and investment.
At the same time, Britain must urgently review the capacity of foreign actors to control its technology stack, and work to both widen its supply base and develop homegrown capacity in strategic areas. We should identify and cultivate British and EU alternatives to US tech majors; by doing so, we can develop functional sovereignty, and resist attempts by any foreign power to circumscribe the commercial choices of our governments and companies. If Amazon Web Services no longer provides cloud hosting for some or all government departments, then British policymakers are less exposed to vagaries of US policymaking.
We must seek leverage over compute, data, and talent, and we must identify and dominate commercial niches that match our pre-existing commercial and technical strengths. The more hyperscale data centres Britain hosts, the bigger the role it can play in the foundational infrastructure of AI, and the more rent it can extract from the tech majors. So too with data; through using the National Data Library, and passing new data trust and data intermediary laws, Britain has an opportunity to become a centre for the exchange and management of data. In talent, Harvard’s trouble is to Britain’s advantage; using US political instability and immigration clampdowns, our universities can poach the world’s leading AI researchers. As for academics, so too for salespeople, product managers, and marketers: the UK should aim to attract the top 100 performers in each category into the UK and place them in UK firms. By doing so, it can supercharge growth.
Economies and societies that host corporate headquarters derive more benefits than branch offices. As a matter of national strategy, Britain should aim to promote and host category-leading AI companies in areas of pre-existing British strength: sectors like law, insurance, finance, pharma, and science. Like Taiwan did over the global chip industry, the UK must also strive to achieve dominance in strategically important sectors; though it may be too late to achieve sustained dominance in the foundation models behind LLMs, we can still look to other areas of the AI and technology value chain.
Above all, we must develop a deeper sense of the national interest. Britain should aspire to be more than a captive market for US tech products, or a declining service economy on the fridges of Europe; we have the talent, the universities and the infrastructure to be a world-leading player. Our path must be that of Meiji Japan, not Mughal India; in transformation lies survival.
a.appcomments {
background: #FFF;
border: 1px solid #446c76 !important;
font-family: benton-sans,sans-serif;
font-weight: 600;
font-style: normal;
font-size: 12px;
color: #446c76 !important;
padding: 12px 50px;
text-transform: uppercase;
letter-spacing: 1.5px;
text-decoration: none;
width: 100%;
display: block;
text-align: center !important;
}
