The Missing Institution: A Global Dividend System for the Age of AI
May 1, 2025
If AI breaks the link between work and income, who gets the upside? This piece explores a future of concentrated AI wealth—and makes the case for a global dividend system to share it. It maps the scenario, the risks, how it could work, and why we need to start planning now.
Anna Yelizarova is part of the Windfall Trust team, and this exploration builds on her recent research.
Sign up to receive future essays from this anthology
A Global Economic Transition We Are Not Ready For
Over the past decade, the conversation around artificial intelligence has been dominated by a familiar set of concerns: breakthroughs in machine learning, competitive pressures between labs, and how to govern rapidly advancing systems that even their creators don’t fully understand. Yet far less attention has been paid to a more profound, systemic question: How will AI transform the economic foundations of society?
No one can predict the future with certainty. But one scenario that deserves far more attention is the possibility that advanced AI could dramatically reduce the role of human labor in the economy. What happens if machines can do an increasing portion of economically valuable work better, faster, and cheaper than people? How would we distribute wealth, maintain social cohesion, and preserve human dignity in a world where labor is no longer the organizing principle of economic life?
This is a more profound disruption than most current economic policy debates are accounting for—but that’s precisely why it requires our attention now.
Importantly, this isn’t some fringe worry. It’s a design goal, embedded in the mission statements of leading AI labs. OpenAI’s charter, for instance, explicitly commits to building systems that “outperform humans at most economically valuable work.” Whether you believe that’s just over the horizon or still decades away, the pursuit of that future is already underway and the fact that it’s being seriously pursued and funded should give us pause. This raises urgent questions that must be addressed before such capabilities are realized.
The countries most exposed to AI-driven disruption are often the least equipped to shape how its benefits and burdens are shared. They lack the tax base to cushion job loss, the institutional infrastructure to adapt, and the geopolitical leverage to demand inclusion. Left unaddressed, this asymmetry could widen existing global inequalities and destabilize already fragile political orders.
Most policy conversations today focus on national fixes: tax reform, regional UBI pilots, retraining programs. But what do we do when the problem is global?
Defining the World We’re Planning For
Too often, debates about AI and the economy break down because of unspoken differences in assumptions—about dominant risks, timelines, or where the disruption will hit hardest. Some imagine a slow wave of job displacement due to diffusion, offset by eventual job creation. Others see a future where human labor is largely obsolete. Still others anticipate a world where AI supercharges productivity but the gains flow disproportionately to the most talented and industrious, while everyone else treads water. These futures demand different responses and without clarity about which one we’re planning for, it can be hard to have a coherent conversation about policy, let alone priorities.
So let’s be specific about the scenario this proposal engages with: it’s a world where advanced AI systems can perform a growing share of economically valuable tasks. Not just drafting emails or debugging code but coordinating logistics, optimizing business strategy, designing drugs, producing media, and autonomously completing complex, multi-week projects—more efficiently and cost-effectively than humans. In this world, labor’s role in value creation shrinks significantly and wages begin to decouple from productivity. The share of income going to workers declines while value accrues with capital owners. Millions find themselves displaced from jobs.
In this world, wage-based incomes could stagnate or fall, even as goods and services become cheaper to produce. If purchasing power falls faster than prices, cheaper goods won’t translate to greater well-being. And so we may find ourselves in a paradox: high productivity, low demand. What happens to supply chains, investment, and economic dynamism when consumer demand falters, not from scarcity, but from exclusion?
At the same time, wealth will begin to concentrate. The firms positioned to automate labor at scale; AI labs, cloud providers, chipmakers, and large corporate users, stand to capture enormous value. Not everyone can become an entrepreneur or license proprietary models. And early glimpses into pricing suggest that access to the most capable systems will remain out of reach for most individuals. The result may be a bifurcated economy, where a minority of firms operate at superhuman efficiency, while the majority struggle with weakened demand and eroding margins.
Meanwhile, governments could face a growing fiscal squeeze. Today’s tax systems rely heavily on income taxes from workers, while often failing to capture corporate profits effectively due to tax loopholes, profit shifting, and relatively low rates on business income. As AI reduces the role of human labor and allows companies to operate with fewer employees, this setup begins to break down. With less income to tax from workers, government revenues decline—just as the demand for public support rises. In many countries, social safety nets are already fragile. Without reform, the gap between what governments can provide and what people need is likely to widen.
This future is also geopolitically asymmetric. Countries home to leading AI labs and digital infrastructure may capture a disproportionate share of the economic value. Others, especially those reliant on cheap human labor or foreign remittances would struggle to adapt.
For decades, offshoring to low-wage countries has been a rational strategy: it maximized margins and minimized labor costs. But in a world where robotics and automation continue to advance, that logic could begin to reverse. If machines become capable of undercutting even the lowest wage floors, it might become more cost-effective for companies to bring production closer to home. That shift wouldn’t just simplify logistics or reduce geopolitical risk. It could undermine the foundations of economic growth in many parts of the Global South, where export-driven industrial jobs have long been central to national development. If those jobs were to disappear without new ones rising in their place, the social and economic consequences could be profound.
This isn’t just a theoretical concern. Many governments in the Global South already face strained budgets and limited capacity to deliver large-scale social support. And yet the economic asymmetry is likely to deepen, even by most conservative accounts. PwC estimates that AI could add $15.7 trillion to global GDP by 2030—but just $1.7 trillion of that is projected to reach the Global South (excluding China). The divide is not just about money but about the tools for resilience, fiscal capacity and institutional readiness.
This scenario shouldn’t be treated as a forecast, and isn’t the only future we should be preparing for. A more decentralized world for example, where access to AGI is widespread and development is diffuse, would bring its own economic dynamics and governance challenges. Some in the ecosystem are focused on preventing extreme centralization altogether, through tools like antitrust law or public AI infrastructure. But steering the future is not the same as preparing for it. The challenge isn’t to bet on one future but rather to map the range of possibilities where our institutional thinking falls short.
Beyond Borders: Distributing Value Through a Global Dividend
As artificial intelligence advances, so does the potential for a profound shift in how economic value is generated. If the rewards of AI accrue mostly to a few companies and countries, we face a troubling prospect: a future in which the technologies transforming global productivity leave vast swaths of the world behind.
One emerging proposal is to build a global institution capable of pooling and distributing AI-generated economic gains across national borders. The vision is early and experimental, but its architecture resembles that of a global sovereign wealth fund: a vehicle to hold and grow economic surplus, not in the name of any one country, but in service of humanity as a whole.
What would such a fund do? One proposed framework is a global dividend system: recurring payouts based on the principle that every person holds a legitimate claim to a share of its value.
The concept may sound abstract, but it’s not without precedent. Alaska’s Permanent Fund Dividend, for example, distributes an annual cash payout to every state resident, funded by the state’s oil revenues. It’s a small-scale but powerful model of how sovereign wealth can be shared directly and universally. The global dividend system proposed here builds on that idea, extending it beyond borders and anchoring it not in geography, but in our shared exposure to the economic upheaval brought by AI.
In practice, that system could start small. A rising global income floor beginning with the poorest communities and expanding over time. This offers one pathway to deploy funds in the near term and help people even before enough wealth for full global coverage is generated. It’s a way to begin building the infrastructure and legitimacy needed for broader distribution later, while offering immediate support where it’s needed most.
In the long run, such a system could ensure that everyone retains the means to meet basic needs in a world where labor is no longer the main vehicle for income. But even in the short term, modest distributions could serve as a stabilizing force and targeting the most exposed could soften the landing and buy time for broader adaptation.
One of the key challenges will be legal: how do we codify the idea that every person has a rightful economic stake in this emerging infrastructure? That will require creative legal thinking to embed personhood-based entitlements into a durable and enforceable institutional framework. This reframes economic inclusion not as charity but as a fundamental human right. A form of economic membership in a shared global system.
Building something like this wouldn’t be easy. It would require legal imagination, international coordination, and public trust. It would likely face political resistance from those reluctant to cede sovereignty, and from industries wary of redistribution. But the idea rests on a principle that deserves more space in our collective imagination: collective stewardship of shared wealth in an era of massive disruption.
Capturing the Windfall: Funding the Global Dividend System
If a global dividend system is to be more than a thought experiment, the question is: where will the money come from?
One of the earliest answers to this came in the form of the Windfall Clause, a proposal by Cullen O’Keefe that proposed AI labs to voluntarily pre-commit to redistributing a share of their profits once they crossed a certain economic threshold—say, 1% of global GDP. The concept remains interesting, but the original formulation is likely insufficient for today’s landscape. The mechanism struggles against familiar headwinds. Profits are pliable. They can be shifted, hidden, reinvested. Thresholds can be gamed or endlessly deferred. If we are to revive or build on the Windfall Clause, it will need to be significantly redesigned: clearer, harder to game, and supported by external levers of accountability.
Even our most established tools, like taxation, struggle to enforce compliance in a globalized economy full of loopholes, havens, and jurisdictional arbitrage. Expecting voluntary commitments to fund a global commons, or for firms to accept supranational obligations, would demand a degree of institutional alignment and political resolve that feels out of reach today. But as the nature of work changes and public expectations shift, what once seemed politically unimaginable may start to look necessary. In a post-labor world, the boundary of feasibility is likely to move.
There are other paths worth exploring. One idea is to secure equity stakes in the firms or infrastructure providers most likely to benefit from AI-led transformation. Unlike profit-sharing pledges, equity provides a legal claim on future value, and can be harder to evade or obscure. But equity-based models assume we know in advance where value will concentrate and that public institutions have the leverage to claim a meaningful stake. In reality, the value chain may splinter across labs, cloud providers, chip manufacturers, and downstream companies quietly automating labor with little public scrutiny.
Another approach might shift focus to the users of AI systems rather than their producers. Imagine a framework where large companies that automate away significant portions of their workforce are required to pay into a global fund as a condition for continued access to advanced AI. But that, too, relies on a fragile foundation: a willingness from AI labs to coordinate, a willingness from firms to comply, and enough shared incentive to keep the system from unraveling the moment it becomes inconvenient.
Then there’s the state. Governments could implement targeted taxes on AI-driven productivity gains or capital income, and route the proceeds into a shared fund via international agreements. It’s a long shot, but in a world where labor markets falter and consumer demand softens, governments might find themselves with incentives to keep markets from collapsing. Economic stability is, after all, a shared interest, even in an unequal world.
There’s no easy, perfect off the shelf mechanism ready to be implemented with enough political will. It is a question that deserves far more attention, experimentation, and collective foresight. But if we’re serious about building a post-labor economy that doesn’t just concentrate power, we’ll need to start designing and testing real, enforceable funding systems.
Recognizing the Limits
No single policy can carry the weight of a future this complex. A global dividend system won’t be enough on its own. It cannot address deeper questions of identity, purpose, and meaning. Those challenges will demand cultural, social, and institutional responses to supplement it. But it can offer a starting point for economic security in an era when the foundations of income and labor are being redefined.
This proposal also assumes a specific kind of future: a centralized world. One in which the economic gains of advanced AI accrue not broadly, but to a handful of companies and countries. That may sound dystopian to many, but that doesn’t mean we should resign ourselves to it. If this future seems undesirable, then it’s urgent we start sketching alternatives, designing institutions and incentives that push us toward better outcomes.
There are other ways to imagine redistribution in this scenario. Some might advocate for universal basic services provided by nations, perhaps funded through multilateral tax treaties. Others might suggest public ownership of key AI infrastructure. None will be easy. Getting governments to tax powerful domestic industries is hard. Getting multinational firms to commit to global contributions is harder. But politics has always been about power, and the exercise of it. If the legitimacy of economic systems begins to fray, perhaps even entrenched interests may discover that redistribution is in their own long-term interest.
This is one proposal, built for one possible future. It isn’t offered as a definitive answer, but as a starting point for debate, refinement, and imagination. The assumptions behind it may prove incomplete or incorrect; only time will tell. But in an era of rapid transformation and deep uncertainty, putting concrete ideas on the table is a way to clarify thinking, surface disagreements, and inspire alternative approaches.
At its best, this proposal is more than a policy suggestion-–it’s an invitation. To imagine a global economy grounded in solidarity and help catalyze discussion about what kind of futures people actually want, and what institutions will be needed to make them viable.
Special thanks to Deric Cheng for his helpful comments and suggestions.
Newsletter
Sign up for our latest articles
We're publishing essays from our expert researchers on the topic of a new social contract bimonthly.