For most of human history, labour was destiny. It built pyramids, laid railways, powered economies and anchored political legitimacy. But the rise of artificial intelligence (AI), with its tireless efficiency and insatiable appetite for tasks, threatens to sever this age-old link. If Bill Gates is right, within a decade humans may no longer be “needed for most things”. It is a startling proposition but increasingly difficult to dismiss.
From boardrooms to basements, the sense of impending upheaval is palpable. Dario Amodei, the CEO of Anthropic, predicts half of all entry-level white-collar jobs could vanish within five years. Mustafa Suleyman, head of AI at Microsoft, is even gloomier, arguing artificial general intelligence (AGI)—a system more capable than humans at economically valuable work—will not merely augment labour, but replace it. Gates himself foresees a world of “free intelligence”, where top-tier medical advice and personalised tutoring are not privileges of the few, but commodities available to all. Liberating, yes. But also deeply destabilising.
AI’s ascent is not hypothetical. It has already surpassed most humans at standardised testing, competitive programming, specialised scientific knowledge, among others . Generative models write code, diagnose disease and ace exams once thought uniquely human. If capital once needed labour to extract value, it may soon prefer a cheaper, more pliable alternative: algorithms that neither sleep nor strike.
The implications are epochal. First to go are white-collar professions: accountants, analysts, paralegals, perhaps even journalists. Blue-collar work, long thought resistant to digitisation, will follow as robotics matures. The economic logic is inescapable: once trained, AI is nearly costless to replicate and deploy. Firms built on margins will have every reason to swap employees for models.
This erodes a foundational contract of modern political economy: that labour grants economic agency and social dignity. The risk is not merely unemployment but obsolescence. If humans no longer contribute meaningfully to production, their bargaining power collapses. This recalls the “resource curse” afflicting oil-rich autocracies: when rulers can fund themselves through petrodollars rather than taxation, they have little incentive to heed their citizens. Replace oil with AGI and the analogy becomes chillingly apt. A world run on artificial intelligence may not need intelligent citizens.
Technocrats offer universal basic income (UBI) as a salve. But paying people to exist is not the same as empowering them to act. UBI may stave off poverty, but it does little to restore meaning or autonomy. The danger is a population economically sedated—fed, housed, but fundamentally excluded.
There is, however, another path. Rather than building AI to supplant humans, why not design it to enhance them? Gates himself points to tools like Cursor, an AI coding assistant that boosts programmer productivity without replacing them. The aim should be a “motorcycle for the mind”, not a bulldozer over it.
This requires rethinking the architecture of AI itself. At present the race for AGI is concentrated in a few firms with extraordinary compute power and troves of proprietary data. If control of AI is monopolised, so too will be its spoils. The alternative is a decentralised model where individuals can train their own models, own their data and run intelligent systems on affordable hardware. In such a world, the worker might resemble a CEO: managing AI workflows, setting strategic direction, commanding machines, rather than competing with them.
Yet even the best tools are toothless without robust institutions. Democracies must act now to embed AI within accountable, inclusive frameworks. This means strengthening governance; enforcing fair rules for access and participation; and resisting the temptation to let elites write the algorithmic rules of tomorrow. The window is closing fast. AI labs are training systems to replace humans. Societies must train systems to serve them.
Gates, ever the optimist-engineer, sees opportunity amidst the uncertainty. He imagines AI curing diseases, mitigating climate change, educating the planet. But even he concedes the transition will be “very profound and a little bit scary”. Rightly so. The industrial revolution disrupted labour markets over a century. The AI revolution may do so in a single decade.
Work, once humanity’s anchor, may soon be cast adrift. What remains is the harder question: not what jobs AI will destroy, but what roles—economic, social, political—humans can still claim in the age of machines. The answer will shape the future of worth. ■
