Work in the Age of AI: Reflections from After Neoliberalism

At last year’s After Neoliberalism conference, questions about how to ensure human flourishing with the development of AI and other emerging technologies took center stage.

Founded on Labor: AI and the Crisis of Purpose in the Post-Neoliberal Era

Last December, the Allen Lab for Democracy Renovation at Harvard convened “After Neoliberalism: From Left to Right,” a conference that brought together more than three hundred scholars, policymakers, journalists, and civic innovators to, in Danielle Allen’s words, “do some thinking together in public.” Allen opened with a reminder that framed everything that followed: “the economy is always for something… and at the end of the day, it should matter for the project of human flourishing.” Across two days and eight panels, one theme kept resurfacing – the American people have become exhausted with the economic status quo of recent decades. Artificial intelligence is certainly not the cause of that exhaustion, which long predates the technology, but it is undoubtedly adding a new dimension of fear and uncertainty to the American state of mind on the economy.

As I listened to the panels at the conference, my mind kept wandering somewhere unexpected: to a seminar room across campus, where this year I took a wonderful class on the Italian constitution. I recognize how random that sounds, as a conference on the future of the American political economy has, on its face, nothing to do with a constitutional text drafted in Rome nearly eighty years ago. But one line from that class had lodged itself in my head, and the conference served as a constant reminder of it.

For context, in 1947, a constituent assembly gathered in an Italy that found itself in ruins. The country had just emerged from twenty years of fascism and a war that left its cities bombed, civic life shattered, and roughly half a million of its citizens dead. The delegates, among them communists, Christian Democrats, socialists, and liberals, agreed on almost nothing. Yet the first sentence of the constitution they produced reads as follows: L’Italia è una Repubblica democratica, fondata sul lavoro. ‘Italy is a democratic Republic, founded on labor.’

Not founded on markets, nor on territory, faith or blood. Of all the values a broken nation might have chosen to rebuild itself upon, a fractious assembly spanning a vast political spectrum converged on the idea that work, or the daily act of contributing to society, is what binds a people into a democracy.

This essay (and this conference) is about not Italy, but what that country’s framers understood, and what the After Neoliberalism conference, taking place an ocean away and many generations later, elucidated so clearly. Work is not merely how people earn; it is how we find belonging. Artificial intelligence is about to test whether work still serves as a bedrock of society. “Work is purpose. It’s not just for consumption,” said Anne-Marie Slaughter, CEO of New America. If Slaughter is correct, then job losses from AI would represent more than an economic crisis. A crisis of purpose may await our country, too.

The Old Model

Policymakers have long operated under a simple premise: if markets are efficient and GDP grows, society will be better off. As Oren Cass of American Compass put it at the conference, the conventional wisdom has long held that if we have more “stuff,” the economy is working. Indeed, there is no denying that for countless people in this country, that model has worked. Americans today enjoy a quality of life that would have been beyond belief to even the wealthiest in generations past. Work, in this traditional framework, is a means to consumption, and technology is inherently good if it produces more output with less input of labor. Distribution has been handled after the fact. In Cass’s words: tax the winners, send a check to the losers, and move on.

AI is demonstrating just how untenable that premise may be. What happens when the “losers” are no longer a small group of displaced workers, but entire categories of labor? What happens if the economy produces never-before-seen abundance but strips millions of people of purpose? Artificial intelligence forces us to ask not only how to make markets efficient in order to grow the pie, but also the more foundational question Allen posed at the conference’s start: What is the economy actually for?

One of the most striking features of the conversations at After Neoliberalism was not ideological conflict, despite speakers drawn deliberately from left and right, but convergence around the urgent need to wrestle seriously with these questions – an echo, one could say, of that postwar assembly in Rome. Panelists agreed that the neoliberal focus on efficiency too often came at the expense of stable communities, meaningful work, and belonging, producing what panelists described as a “disconnection crisis.” During the opening panel, Slaughter, who also served as a senior official in President Obama’s State Department, emphasized that humans are not “rational automatons” but social beings who need connection and purpose. Aaron Hedlund, an economist at the White House Council of Economic Advisers, called for a shift from “utilitarian consumerism” and “unbridled efficiency” to “productive dignity and purpose.” As Mr. Hedlund put it, “labor is not some divorced thing from capital. Human capital is actually a very essential form of capital.”

A Check, or a Job?

Can a check substitute for a job? Few make a stronger case for the answer being ‘no’ than Cass. In his fireside conversation with Rebecca Henderson, Cass argued that neoliberalism’s signature move, “taxing winners and sending losers a check,” fails not because the checks are necessarily too small, but because it misunderstands what people want and need in their lives. First and foremost, for those physically and mentally capable of working, it removes the dignity of being contributors to one’s community and society, rather than recipients. As Cass put it plainly: “The social safety net cannot replace a lack of decently paid jobs.”

Even worse, argued Cass, the current system actively disincentivizes the very thing it should prize, which is work. Benefits structured around non-work penalize those who take low-wage jobs, devaluing and disincentivizing the very thing that should form part of the foundation of people’s lives.

The remedies Cass outlined were strikingly concrete. He advocates for replacing portions of traditional welfare with a wage subsidy, support attached directly to the paycheck, so that low-wage work becomes sustainable and every additional hour worked is rewarded, rather than clawed back. He wants to restructure the safety net to support those who cannot work generously, while aggressively rewarding those who can and do. He believes that government should pursue tight labor markets, even a deliberately engineered “labor shortage,” so that employers must innovate and raise wages to attract workers, rather than treating labor as a cost that can be squeezed.

These are policies founded, quite literally, on labor. A check may help pay the bills, but it offers nothing in the way of meaning, connection, or community.

“Pro-Worker AI”

Any debate on AI’s role in the economy should, ideally, be rooted in the impact it will have on working people.

The pessimistic scenario is familiar: AI automates tasks, firms cut labor costs, and wealth concentrates in the hands of the few who own the technology. A dystopian future emerges in which work is more precarious and hollow than it already is for many today. The service sector is gutted, and as one panelist on the technology panel warned, human contact could become a premium add-on. Consumers may soon have to “plus up” to talk to a real person, with a chatbot serving as the new default.

The conference pointed to another path America could choose to take. Daron Acemoglu, the MIT economist whose work has profoundly shaped scholars’ thinking on technology and labor, argued for what he called “pro-worker AI” – technologies that “expand what workers can do” rather than automating them out of existence. Acemoglu insisted that this path is both “technically feasible and socially desirable.” Henderson made a similar point in her conversation with Cass: AI must be a “complement, not a substitute” to labor. Instead of replacing a nurse, AI could lessen the administrative burden and free up time for hands-on, patient-centered care. Instead of automating instructors, it could facilitate access to personalized instruction and learning while allowing teachers to focus on mentorship and personal connection.

Under neoliberalism, technology has been treated as an exogenous force that can be adapted to but not shaped. The conference panelists broadly rejected that fatalism, with participants arguing time and again that technology is inherently political, considering that algorithms, data, and deployment incentives all reflect human choices. As one panelist on the technology-focused panel insisted, “We must be the master of the technology, rather than the other way around.”

Acemoglu warned that right now, the trend is towards centralization of data, computing power, and decision-making authority, risking not just inequality, but a new digital and societal oligarchy in which the infrastructure of knowledge itself is concentrated in the hands of a small few. In that scenario, owners of capital would grow wealthier and more powerful, while the precarity facing workers, both in our economy and in our democracy, risks being exacerbated. But as Erica Smiley noted optimistically, workers are well-positioned on the front-lines of the fight for democracy. She argued that working people are already fighting for “self-determination” on a range of issues within the workplace, such as better safety conditions and decent wages, proving that democracy is “alive and raging” in “one of the frontline battlefields: the shop floor.” If the workplace serves as the frontline of the battle for democracy, then a warning offered by Henderson becomes all the more poignant: “workplaces cannot be sites of authoritarianism.”

Care: The Work AI Cannot Do

Perhaps nowhere are the stakes of the American economic transformation clearer than in care. “Everyone will be a caregiver at some point in their lives,” Slaughter observed, and care work, though economically essential, has long been undervalued. Allen said that the care economy is “at the center of how work is structured, and therefore at the center of the economy” itself, and panelists were in broad agreement that the health of the family more broadly should be a critical measure of economic success.

Nevertheless, the neoliberal era saw the value of care cheapened. Political scientist Alisha Holland noted that the rise in dual-income households facilitated women’s entry into the workforce and simultaneously “deactivated” the traditional, extended family care network. Brad Littlejohn of American Compass argued that the “two-income trap” robbed parents of the “flexibility” to stay home and care for children or elders.

Yet perhaps paradoxically, the much-lamented shortage of care workers is, according to economist Dani Rodrik, only a myth of framing. These are not unfillable jobs, but “terrible, terrible, terrible jobs” offering low pay and subpar conditions that, as a result, must be filled largely by foreign workers. Rodrik argued not for an end to such immigration, but rather for “upgrading and upskilling” – improving working standards, raising wages, and using new technologies to ease the burden on workers.

Across the two days of the conference, panelists suggested a host of other solutions to the crisis in the care economy as well as innovative alternatives to how care is handled in the United States. Rodrik highlighted the “Neighborhood Care Groups” model from the Netherlands called Buurtzorg as an alternative to private equity-driven care. Holland pointed to the “care block” system in Bogota, where the city government organizes support for caregivers, such as providing care for children while their caretaker goes to the doctor. In an era of backlash to mass migration, she also discussed the potential for “circular short-term migration” and community-driven immigration, where local towns could sponsor the specific workers they need, including care providers.

While AI may assist in the care domain too, it fundamentally lacks empathy and intention, and so the need for humans is unlikely to dissipate. Even when an algorithm delivers a service more efficiently, it cannot replace the power of the metaphorical – and in the case of the care economy, the literal – human touch.

The same holds across the service sector, where many of the panelists predicted the jobs of the future will actually lie. As machines and AI absorb routine tasks, the human elements of labor, such as empathy, judgment, and connection, become more valuable, rather than less. The danger, however, is that markets will fail to reward those qualities, just as they have too often underpriced care. The opportunity now is to build an economy that rewards workers in the sectors in which the role of the human being is foundational and irreplaceable. Ironically, these may be precisely the sectors, such as care, that neoliberalism long neglected.

What We Are Founded On

The Italian framers of 1947 were not sentimental about work. They had watched a regime promise national glory while stripping people of their basic rights, and so they concluded that a democracy is only as sturdy as its citizens’ capacity to contribute to it. Labor was their answer to fascism: a republic of contributors cannot easily be ruled by a demagogue.

The conference arrived at a similar place. “Necessitous men are not free,” Felicia Wong of the Roosevelt Institute reminded the room, invoking Franklin Roosevelt’s 1944 State of the Union — the speech in which FDR proposed a Second Bill of Rights and declared that true individual freedom cannot exist without economic security. Yet democracy requires an economy that provides both security and purpose. As legal scholar Jed Purdy succinctly said, the modern social safety net focuses on pure financial support, not access to good jobs. A society in which AI generates immense wealth but leaves large portions of the population without meaningful contributing roles is not a success, even if living standards broadly rise, because transfers cannot substitute for purpose. Without a job, said Luma Simms, of the Ethics and Public Policy Center, we lack the “number one need for the human soul: rootedness.”

After all, the AI debate is not really about technology. It is about our values as a country. Do we want an economy that treats people as costs to be minimized, or contributors to be empowered? Do we accept a future in which human interaction evolves into a luxury good? Will a handful of firms fully shape the architecture of artificial intelligence, or will America build systems that reflect our commitment to both capitalism and democracy? If transfers – a default solution that creates dependencies, leaves people unable to stand up to the owners of capital, and potentially disincentivizes work – are not the answer, is government up to the enormous task of preserving access to decent-paying jobs?

AI will not answer these questions for us, but it will make them impossible to avoid. The neoliberal era assumed that if the economics were done right, the rest would follow. The lesson of this moment is the opposite: a democracy must decide what it is founded on.

Charlie Covit is an undergraduate student at Harvard College and Researcher at the Allen Lab for Democracy Renovation.

The views expressed in this article are those of the author alone and do not necessarily represent the positions of the Ash Center or its affiliates.