The Age of the Bargain: How Societies Will Survive the AI Revolution

I have never written a prediction before. As a teacher and learning designer, I usually concern myself with how people learn, not with forecasting the future. But over the past year, it has become increasingly hard to ignore what is happening around me. The fields I work in — education, writing, learning design — are being rewritten by artificial intelligence. Tasks that once took hours of human judgment now take minutes. What once required teams of professionals can often be done by a single person with the right prompt.

It is not just an efficiency story. It is a civilizational one.

I keep wondering what this means for the next generation — students and young professionals stepping into a world that may no longer need their labour in the way it once did. They are taught to prepare for work, but what if work itself is being redefined faster than we can adapt? What happens to a society when the link between effort and livelihood begins to fray?

I have never felt less confident about the timeline of change, but more convinced about its direction. The future we are entering is unlikely to be a sudden dystopia or a technological utopia. It will be something slower, more chaotic, and human: a turbulent transition in which societies struggle to strike a new bargain between capital, labour, and the state.

The first wave of disruption is already here. Corporate behaviour offers the clearest clue. Under pressure from investors to stay competitive, companies are deploying AI not for innovation in the abstract, but for ruthless efficiency. The signs are subtle but unmistakable: hiring freezes in entry-level roles, shadow layoffs in tech and consulting, and a renewed pride in “doing more with less.” Tasks once performed by junior analysts, copywriters, and customer service agents are being absorbed by automated systems. As Ernest Hemingway once wrote about bankruptcy — "gradually, then suddenly" — that is how the AI revolution is likely to unfold.

This is the asymmetric shock of the AI age. Where most of the tasks in a role can be absorbed by machines, the role itself thins out or vanishes. Where AI handles only parts of the work, productivity rises, and expectations rise with it. Some workers find their capabilities amplified; others discover that their contribution has become optional. Augmentation and replacement coexist, but they distribute risk and reward very unevenly.

For a time, the economic data may even look deceptively strong. Productivity increases. Corporate profits grow. But beneath the surface, social strain accumulates. Wage stagnation for many and windfall gains for a few deepen existing divides. The psychological toll mounts quietly: a sense of redundancy among those still employed, and resentment among those who are not. The system heats up long before policymakers register that it is overheating.

At this stage, many observers reach for a reassuring counter-narrative: that AI is mostly hype, a speculative bubble destined to burst. There may be truth in the financial diagnosis. Investment has raced ahead of revenues, and history is littered with overbuilt futures.

But even if valuations collapse, the underlying shift does not reverse. Organisations do not forget what automation has taught them. Past bubbles — railways, electrification, the early internet — changed the pace of adoption, not its direction. A bursting bubble may slow hiring and bankrupt weaker players, but it does not resurrect vanished roles or restore the old correspondence between effort and livelihood.

At some point, the social cost of this first phase becomes impossible to ignore. An economy cannot thrive indefinitely if large parts of the population lack security, purchasing power, or a sense of stake. The question “Who will buy our stuff?” becomes uncomfortably real.

Some far-sighted business leaders begin to advocate for intervention, not from altruism but from self-preservation. Governments, under pressure, respond as they usually do: late, improvisationally, and with imperfect tools. What follows is not smart governance, but crisis management.

The result is an uneasy compromise, a new social bargain born of necessity. Governments expand the safety net not in one sweeping move, but in layers: enhanced unemployment insurance, targeted cash transfers, tax credits, food and housing support. Together, these function as a partial basic income for large sections of the population. To fund them, states reach for the only politically viable lever: taxing the winners. Robot taxes, higher corporate rates, and special levies on AI-driven profits appear, not to halt automation, but to recycle part of its windfall into social stability.

Parallel efforts attempt to regulate the worst abuses — bias audits, limits on automated firings, data-use restrictions — unevenly enforced and riddled with loopholes. Over time, these measures coalesce into what I call the Bargain: the implicit deal by which elites accept limited taxation and oversight in exchange for social stability and the preservation of control.

The economist Yanis Varoufakis has described the emerging order as “technofeudalism”, dominated not by industrial capitalists but by digital landlords who own the platforms through which others must live and work. The Bargain is society’s attempt to negotiate with these new overlords rather than overthrow them. It is not justice; it is triage. But it works, at least for a time.

Beyond the backlash lies the slow settling of a new world. By the late 2030s or 2040s, advanced economies adapt to the AI transformation not through brilliance, but through weary pragmatism. A new social structure solidifies, defined less by occupation than by access to assets and algorithms.

A small, immensely wealthy elite owns the AI systems, data, and infrastructure that drive the economy. Beneath them stands a smaller tier of workers whose skills complement these systems — well compensated but permanently precarious. Below them lies the largest and most unstable group: people whose livelihoods depend partly on an expanded welfare state and partly on gig-like, AI-mediated work. They are not unemployed, but underemployed, moving between short-term contracts, micro-jobs, and digital hustles.

For this group, economic precarity coexists with digital abundance. Entertainment, simulation, and algorithmic distraction are plentiful. What grows scarce is dignity rooted in work. As work loses its role as a primary source of identity, the search for meaning migrates elsewhere: online communities, political movements, caregiving, creative expression.

There is nothing inevitable about this scenario. But it is plausible because it aligns with the enduring logics of power, profit, and adaptation. Every major technological upheaval began with dislocation before new forms of stability emerged. The idea that this transition will be faster or smoother misunderstands how complex societies evolve.

Nor does this scenario depend on moral awakenings. It assumes only that institutions, when pressed hard enough, act to preserve themselves. They always have.

The central conflict of the twenty-first century will not be between humans and machines, but among humans themselves — over the terms of the Bargain. The question is whether a society without universal work can still offer dignity, purpose, and community. Bread and circuses can pacify, but they cannot fulfil.

The age of the Bargain is not a distant future. It is already beginning around us, in invisible negotiations between progress and justice, efficiency and empathy. The outcome will depend less on the machines we build than on the humans we choose to be.