If AI Takes the Work, What Goes With It?
San Francisco, March 2026 – University of Notre Dame's Institute for Ethics and the Common Good, The Brookings Institution
Our Scenarios Program convenes economists, AI researchers, policymakers and civil society leaders to explore the implications of transformative AI, and to identify the policy responses each future demands. The workshops are complemented by the Policy Atlas—a continuously updated guide to policy options.
We believe that countries should develop Economic Preparedness Plans. Our workshops and the Atlas are designed to provide the foundation for doing so.
The conversation about AI and work has, until recently, been almost entirely about numbers: how many jobs, in which sectors, how fast. That conversation matters. But it skips a prior, more foundational question—what work is actually for. What does a job give a person beyond a wage, and what would we have to rebuild if the machines can do the work instead?
That question brought an intimate group of philosophers, theologians, social scientists, AI researchers, and policy experts to San Francisco on 17 March 2026. The workshop was co-convened by the University of Notre Dame's Institute for Ethics and the Common Good, the Brookings Institution, and the Windfall Trust, around a shared conviction: the debate about AI and work is missing something essential.
One view, voiced loudly and confidently in Silicon Valley, is that AI will free humanity from the "servitude" of work and that meaning will take care of itself. Because the people who hold that view are the ones building the systems, this assumption—if left uncontested—risks being designed into the technology and the policy that follow, settled less by argument than by inattention.
We held the day under the Chatham House Rule, so what follows reflects the themes that emerged rather than the views of any individual. We met in San Francisco deliberately—to immerse ourselves in the milieu of companies building this technology, and then widen the lens.
The core challenge: What is work for?
Do we truly know what’s at stake in losing work? Economists often describe work as a bundle of goods that happen to arrive together—income, yes, but also social belonging, daily structure, a sense of competence, recognition, and the feeling of contributing to something beyond oneself.
If work is only a delivery mechanism for separable goods, then in principle each one could be supplied some other way—income through transfers, social belonging through other institutions, and so on. But can it? Or does something essential, something like dignity, disappear in the unbundling?
This was the question at the centre of the convening—and the answer isn’t merely existential, it has concrete policy implications. If dignity is inseparable from economic contribution, then does preserving meaningful work—even at real cost to efficiency—become a moral imperative? If instead, dignity can be sustained through other institutional forms, then should the task be to build the institutions that can deliver it? Either way, you cannot choose a policy path without first taking a position on the question.
Two scenarios, neither reassuring
To make the question concrete, participants worked inside two near-future scenarios, offered as thought experiments rather than predictions.
In the first, The Post-Work Society, general-purpose AI performs nearly all cognitive and creative labor, and societies choose to move away from work as the organising principle of economic life. Abundance is shared through high taxation of AI-driven profits, universal basic income, and a "public AI" infrastructure treated like a utility. Work becomes optional; human labor survives in niches where empathy, embodiment, and creativity are prized.
In the second, The Great Pushback, AI reaches human-level capability but adoption is uneven. Licensed professions—medicine, law, engineering—slow deployment through oversight and liability rules, while less protected white-collar workers face sudden displacement and backlash. Societies respond by deliberately preserving jobs—human-in-the-loop mandates, a "right to human review," shorter work weeks, retraining. The labor market tilts toward care, service, and direct human contact.
What was striking was that neither path reassured the room. The post-work vision drew particular scepticism—not because abundance is unwelcome, but because the assumption that freed-up people will simply find meaning, and that such a transition would unfold smoothly, looked like exactly the kind of forecast that tends to fail: the ones that assume people will become different from what they are. The pushback scenario carried its own warning, too—preserving employment can shade into hollowing it out, with people reduced to supervising machines in dehumanising roles redesigned around the technology rather than around them.
What we saw
Several threads ran through the discussion.
The real threat is to meaning, not just to employment. Loss of income is the part for which we already have plenty of potential policy responses. The other losses—structure, identity, recognition, the sense of contributing—are harder to compensate for. A period of mass dislocation is a precarious moment—when people feel lost, they are vulnerable to those who would exploit them, yet it is also a moment when communities can come together in ways not otherwise available. How societies "companion" people through this period, toward something generative rather than despairing, may matter as much as any transfer payment. What's needed, one participant suggested, are “transformation doulas”—people to accompany others through the change, not simply to compensate them for it.
Meaning will not organise itself. There was firm resistance to treating faith, philosophy, and purpose as palliative care to be administered after the economic questions are settled. If reflection and discernment matter in an AI-shaped future, they matter now, while the trajectory is still being set. As one participant put it, the frontier labs are already "in the meaning business," whether they intend to be or not. The implication was that the people who think about meaning for a living should be present while the technology is designed, not called in afterwards to console those it displaces.
"Dignity of work" needs sharpening. Participants asked whether "dignity" lands across political and cultural lines, or whether the ideas need different framing to carry. A distinction emerged between the objective value of work—what it produces, where advanced AI may eventually outperform us—and its subjective value—the way work forms and develops us as people, and through which we contribute to a common good. The first can be automated away; the second is where much of the dignity lives, and it does not transfer to the machine that takes over the task. Participants also pushed to widen "work" beyond waged labor to include care and family work, and the other unwaged activities that give life meaning.
Preserving jobs is not the same as preserving their worth. Even where work survives, reorganising it around AI can degrade it. The industrial-revolution echo was explicit—people had to become machine-like to work beside machines, and with AI that risk returns in a new form.
The conversation is missing the people it's about. The discussion kept returning to who was not in the room—workers themselves, the wider public, and people beyond the wealthy post-industrial world. For the Global South in particular, access to AI-driven abundance could be transformative rather than threatening. The same change looks very different depending on where you stand, and frameworks built without those voices will misjudge how the change is actually felt.
What comes next
This San Francisco gathering was the first phase of a year-long effort led by Notre Dame's Institute for Ethics and the Common Good, Brookings, and the Windfall Trust. Its job was to surface the questions and cruxes worth exploring in depth. A second phase will convene a wider, deliberately cross-partisan group of public intellectuals in Washington, D.C., turning to questions of more immediate policy relevance. A third will take the work public, through writing, events, and media.
To the Silicon Valley optimists, unbundling the intangible rewards of work from the wage means shedding the drudgery while keeping everything that gives work meaning. Many at the workshop doubted whether it splits so cleanly. Structure, recognition, and a sense of contribution might survive the separation, but not on their own, and not without deliberate, inclusive design and the institutions to provide them. We are running that experiment now anyway, on everyone, and mostly by inattention. Whether dignity survives the unbundling is still ours to decide, for now.
Reach us at contact@windfalltrust.org, or follow the work through our newsletter, the AGI Social Contract, at windfalltrust.substack.com.
