One of the most important and uncertain debates on artificial intelligence centers on whether AI will lead not just to jarring job disruption for millions of workers but to an actual marginalization of humans in the workforce. A world without work—or with far less work—as we know it.

There is growing consensus that AI could lead to dramatic job dislocation. The International Monetary Fund’s Kristalina Georgieva has warned of a “tsunami hitting the labor market.” Yet many, from Goldman Sachs’s Jan Hatzius to McKinsey to former Google CEO Eric Schmidt, argue that AI will follow past technological changes and add net jobs and growth. Schmidt suggests that even if we initially see, for example, “the vast majority of programmers… replaced by AI,” overall “there’ll be more jobs created than jobs lost.” But many top academics and technologists argue that “this time will be different”; that AI is different. This time, technological change will lead to a marginalization of humans in the workforce and a net reduction of jobs. Bill Gates projects that in 10 years, humans will not be needed “for most things.” Anthropic’s Dario Amodei predicts a world in which “the economy grows at 10 percent a year, the budget is balanced—and 20 percent of people don’t have jobs.” Yuval Noah Harari warned in 2017 that “we might witness the creation of a massive new unworking class: people devoid of any economic, political, or even artistic value.”

Offering policy frameworks for an AI era means dealing with exceptional uncertainty. On the one hand, some argue there has been little AI-related job loss so far while others remind us of false projections made a decade ago that by now radiologists would be obsolete. On the other hand, no policy will be sweeping enough for those anticipating the quick arrival of an Artificial General Intelligence (AGI) that can outperform humans across most tasks. Despite this uncertainty, the stakes are too high for a wait-and-see approach. Now is the time to be developing policies that move in the right direction and can be modulated or accelerated to respond to the depth and speed of job and economic disruption.

Top experts have recently suggested that academics and policymakers start developing policies under an assumption of “transformative artificial intelligence” (TAI)—disruption greater than was caused by the Industrial Revolution, over a shorter time period. Without adopting the precise formulation of TAI, this essay takes the approach that it is both wise and prudent to start devising policies that respond to potential scenarios in which AI does not end most human work but leads to unprecedented job dislocation and, left to its own, at least some long-term reduction in net demand for human work.

This essay rejects the idea that the best path for the AI era is one that gives up—or is even neutral—on the imperative of work and human contribution as fundamental to our economy and human dignity. National policy should proudly stake out a “people with purpose preference.” We should be affirmatively steering the use of AI toward augmenting jobs and adding to their quality and value. And where AI leads to the loss of millions of jobs or even job categories, we should be pushing policy to create new jobs and broader forms of economic contribution that promote the well-being of all our people.

The approach here resists utopian assumptions of unlimited resources or models like Universal Basic Income (UBI), Universal Basic Capital (UBC), and mechanisms based on the Alaska Permanent Fund that would seek to exhaust any revenue raised from AI productivity gains through equal distributions to all, regardless of need, circumstance, or what current economic dignity harms would go unremedied. Instead, we should have a more purposeful and targeted Economic Dignity response: A three-part framework that dedicates such AI funds to: 1) initiatives that spur massive new job creation in areas that provide purpose and meaning to both workers and those being helped, 2) targeted efforts to ensure a basic floor of economic dignity for anyone engaged in any type of work or economic contribution, and 3) a national commitment to not repeat past errors of harshly leaving behind segments of our people even when such disruptive economic change is accompanied by net job creation and national growth.

I. Shaping Our Economic Destiny Requires Clarity on Our Destination

Today it is essential to recall what Erik Brynjolfsson and Andrew McAfee wrote inThe Second Machine Age: “Technology is not destiny. We shape our destiny.”

The first step in shaping our destiny must be knowing our destination. That means knowing our desired end goal for our economy in terms of lifting human well-being in a new AI era. That question is no doubt subjective. DeepMind’s Demis Hassabis argues that addressing the implications of AI is going to require “new great philosophers.” No matter how subjective or philosophical this inquiry is, it is unavoidable. You cannot rationally plot a policy roadmap without knowing where you want to end up. For example, if AI has the potential to lead to more systemic joblessness and growing inequality, what is our end policy goal? To ensure subsistence? Maintain the status quo? Or should it be something more aspirational that elevates human well-being in an AI era? The AI era makes it even clearer that metrics like GDP and productivity growth have never been ends unto themselves but are metrics whose value lies in the degree to which they lift human well-being, happiness, and dignity across the board. Indeed, some of the most doomsaying scenarios of AI’s impact on jobs, workers, and inequality assume unprecedented GDP and productivity growth.

In early 2019 in Democracy, I took my shot at answering the existential question of what our nation should consider our “end goal” for economic policy, including how that should shape our approach to AI disruption. My answer was a three-part definition of economic dignity. 1) The capacity for a worker to both provide security and opportunity for their family, and to be there for life’s most universally meaningful experiences with loved ones. 2) The guarantee that neither the accident of birth—nor the many accidents of life—block the chance to pursue potential, purpose, and meaning in the economy. 3) And finally, the right to work with respect and with freedom from domination and humiliation.

The need for clarity on our ultimate economic end goal has never been more essential. The Trump White House and many of its most powerful corporate allies often treat efforts to ensure AI serves American values such as work, liberty, and dignity simply as obstacles to an end goal of furious GDP growth and AI-induced wealth creation—often reflexively justified by the need to outrace China.

In that framework, economic dignity is at best treated as an afterthought, something to deal with once trillions are made and the war is won. As a nation, our North Star must always be the economic dignity of our workers and families, and we must make sure AI development and deployment moves in that direction.

 II. Universal Economic Dignity, Not Universal Basic Income

Clarity on a national end goal of economic dignity in an AI economy is critical to charting the direction of progressive policy. While an economic dignity goal cannot just be inserted into an AI chatbot for a detailed legislative road map, it can provide a guide. I addressed the possibility of a “this time will be different” AI scenario in my book Economic Dignity six years ago. Then, as now, I felt that policy responses like UBI and UBC profoundly fail an economic dignity goal on three grounds.

First, they fail to recognize both the importance of work that has meaning, purpose, social connection, and contribution, and the degree to which so much of the harm to economic dignity in our nation could be remediated through millions of additional jobs aimed at reducing pain, poverty, and missed opportunity. This dual blind spot leads them to miss something profound: the potential to respond to reduced need for work in a future AI economy by creating millions of new, high-quality jobs that address economic dignity denials in our current economy.

I call these the missing “double dignity jobs”: dignified work that provides economic dignity to others. Double dignity jobs provide workers the dignity that comes with adequate pay, benefits, leave, voice and career paths, while those workers help provide dignity to those in need of vital care, first and second chances, and various forms of assistance. Their work, in turn, unlocks a virtuous cycle, allowing even more Americans to pursue purpose, potential, and economic security for their loved ones.

Second, UBI and UBC provide a misguided concept of “universality.” UBI is “universal” in that it gives everyone—regardless of need or circumstances—the same amount each year. UBC does the same but with capital. An end goal of economic dignity, on the other hand, focuses on a “universal” basic floor of economic dignity for all Americans regardless of the accident of birth, accidents of life, or downturns or tumultuous disruptions in the economy. It focuses not on thinly spreading payments to everyone, but on targeting resources to fill specific gaps in an economic dignity floor. These gaps include inadequate wages, unaffordable care, lack of paid leave, and insufficient job and educational opportunities for workers and their families. Consider health care. An economic dignity approach to universal health is the guarantee that no one is denied affordable care when it is desperately needed to care for loved ones, not providing an equal payment to everyone each year regardless of health status. UBI and UBC advocates may argue that the dramatically larger costs that emanate from universal payments expand the political support for such programs. Yet extending benefits up the income scale to more middle-class families—as opposed to discarding all needs-based conditions—can buy much of the same political resilience without crowding out spending in other vitally important policies that enhance economic dignity. Indeed, the unprecedented anxiety over the future of jobs and economic security in a transformative AI scenario is likely to dramatically expand the number of Americans who will feel a strong self-interest in a stronger, more universal economic dignity floor for themselves and their children.

Finally, UBI and UBC fail to prioritize policies designed to prevent economic dignity harms tied to tumultuous economic change even where net jobs are expanding on a national level. From the Industrial Revolution to the devastating downward cycles of factory towns hit by expanded global trade, public policy has spectacularly failed to prevent deep scarring of segments of workers and communities left behind. Policies that appear to benefit other parts of the nation—and particularly elites—while leading to long-term unemployment, deaths of despair, and devastated hometowns for some cut deeply against any sense of a universal commitment to economic dignity. As we may be on the precipice of the mother of all job disruptions, a framework for addressing unprecedented job loss caused by AI must prioritize how dislocated workers find new paths to meaningful work.

A UBI or UBC response to transformative AI must be judged by comparing its benefits against the economic dignity framework it crowds out. UBI proposals like Andrew Yang’s Freedom Dividend cost over $3 trillion a year. A more recent proposal by Sam Altman through an American Equity Fund would cost about $3.6 trillion annually. Certainly, many would benefit from universal payments each year. But the question must be, how do such payments hold up against what could be done if we had a more targeted approach to create a universal guarantee of economic dignity? Just as an example: Our nation could fund or partially subsidize 20 million high-purpose jobs serving others at a cost of $75,000 each, a $10,000 childcare/child allowance payment to all working families with younger children, $10,000 in wage bonuses, and wage insurance or training support for 50 million people a year and still save roughly $1 trillion compared to UBI proposals. These trade-offs must be weighed.

III. A Three-Part Economic Dignity Framework for an AI Era

A three-part economic dignity frame for AI would: 1) create millions of double dignity jobs; 2) create a universal economic dignity floor with targeted interventions in wages, health care, paid leave, and savings; and 3) expand economic adjustment policy to ensure that even where technological progress has macroeconomic benefits, it does not create severe scarring and dignity harm to any portion of our nation.

Significant uncertainty over the speed and impact of AI makes detailed policy prescriptions difficult. But below, I do my best to lay out how this three-part framework can serve as a North Star to guide progressive policymaking.

Part 1: Double Dignity Jobs

We Currently Face a Major Double Dignity Jobs Deficit

UBI and UBC proponents generally assume that if we did face a reduction in the need for work, we would then have no use for the work and contributions of millions of our people. That’s wrong. We should not cast aside the belief that, as Bill Clinton once said, “we don’t have a person to waste.” Today, if we did a society-wide audit of the jobs needed to ensure economic dignity, we would find a jobs deficit of millions and millions of double dignity jobs, missing only due to lack of resources.

First, consider care jobs. In the United States, caring for a child or young adult with a serious physical, intellectual, or developmental disability is often a near-full-time job. Yet, of the 8 to 16 million Americans with disabilities like autism, Down syndrome, or cerebral palsy, only a small fraction receive needed individualized assistance. This is largely because a lack of public resources leads to only about 1.5 million skilled direct support professionals (DSPs), and virtually none with adequate compensation. The lack of double dignity jobs contributes to the fact that only about two out of five people the Bureau of Labor Statistics counts as having disabilities are in the workforce, despite the vast majority wanting to work. It also forces a significant number of parents to pull back on work and careers or face the exceptional stress of balancing work with heavy care responsibilities, yet this figure does not even include the ever-increasing need for millions of families to find skilled home care for those with Alzheimer’s or other forms of dementia.

Second, consider the jobs needed if we take seriously the American ethos that no one should be denied a shot at pursuing their potential due to the accident of their birth. How many skilled educational professionals would we need to add to the workforce to unlock the potential of every child, regardless of income, disability, or life circumstance? While young people in elite private schools have one counselor for every 20–50 students (on top of private admissions services), missing double dignity jobs means that 8 million students are in schools without a single counselor. In 12 states, there’s fewer than one counselor for every 400 students. These are missing double dignity jobs.

And what about the mental health needs of our young people? As we face a teen mental health crisis, there is only one school psychologist for every 1,127 students in K-12 schools nationally, while some states have only one counselor per several thousand students.

Third, consider where serious intervention by skilled professionals could help halt downward cycles many face through addiction, long-term unemployment, mental health challenges, or the struggle to productively re-enter the economy after being incarcerated.

These are just a slice of double dignity jobs. In Economic Dignity, I outlined six categories of these jobs, all of which have shortages amounting to millions of missing, purposeful jobs: childcare workers and early education teachers, college advisors, jobs to empower those with an intellectual or developmental disability, job navigators for those recently released from prison, or coming back from addiction or long-term unemployment as well as psychologists focused on the whole child, and eldercare professionals.

Thus, a key part of the answer to the potential for AI to reduce the number of current jobs is not to accept massive joblessness and then transfer AI wealth to everyone regardless of need. It is to use the excess wealth created by the AI productivity boom to fund millions of missing “double dignity jobs.”

Double Dignity Jobs Can Be Expanded and Augmented in an AI World

Arguing for an expansion of double dignity jobs should not be seen as standing on the beach trying to keep the AI tide from coming in. These jobs can prosper in an AI era for three reasons.

First, many leading economists project that jobs requiring a human touch and spontaneous problem solving will be resilient in an AI era. While some AI futurists tout a world in which AI agents and robots will make even care workers, coaches, and educators a thing of the past, few of these techno-optimists have any special competence in the degree to which we are—to paraphrase Barbra Streisand in Funny Girl—still “people who need people.” For many of us, the wonders of AI are unlikely to completely override the desire for an actual supportive coach, teacher, tutor, navigator, dementia care worker, early childhood teacher, or caregiver who hugs, smiles, sympathizes, and encourages—even if future Alexas can make similar sounds. That’s why even Sam Altman, who lauds the medical diagnostic prowess of ChatGPT, still says, “I really do not want to … entrust my medical fate to ChatGPT with no human doctor in the loop.”

Second, there is widespread optimism that many of these kinds of jobs can be augmented, not replaced, by AI. Some experts see AI as not replacing health professionals, but reducing burn-out that leads to health care job shortages while empowering them with more time to “delve deeper into the socioeconomic and psychological determinants of health and expand the doctor-patient relationship.”

While some buy into the idea of AI replacing teachers, most see AI increasing the value of educators by enabling individualized coaching, allowing students to move at their own pace without tracking students in a way that is stigmatizing. Notably, the leaders of online individualized education agree that AI will enhance, not replace, the role of teachers. Sal Khan, CEO of Khan Academy, sees AI helping to “support the teacher, inform the teacher, help them with lesson planning and grading, so that the teacher can elevate and do more of the human element of things.” Coursera’s CEO, Jeff Maggioncalda, sees AI “not as a replacement for educators but as a powerful tool to enhance their capabilities and reshape the way we approach learning at scale.” My mother was a true pioneer in the sphere of individualized learning—but having to manually judge each student’s work separately made her work insane hours. Educational experts see the potential for AI to facilitate teachers becoming super-teachers and super-coaches without such punishing hours.

Investment in Double Dignity Jobs Can Generate High National Returns

A double dignity jobs approach can rectify our failure to recognize the fact that key public investment can bring future national savings and thus high national returns on investment. In the private sector, such future returns can be instantly calculated into a company’s net present value. With public investment, returns often come over long periods of time, with the savings coming in differing government programs in varying levels of federal, state, and local government. Currently, we lack a national ledger that allows returns on public investment to be credited to the unit of government making the investment. We know that investment in quality pre-school can return $7–$10 for each $1 invested, yet the $1 is simply logged as a cost with no return on investment. If we are creating an unprecedented national program to use new revenue streams from AI specifically to ensure national well-being, it makes sense to consider the degree to which many double dignity jobs can have high returns and even pay for themselves by reducing future costs of hospitalization, addiction, incarceration, or long-term unemployment. Indeed, the potential for AI to augment double dignity jobs that reduce future costs and increase labor force participation further justifies funding for high-ROI double dignity jobs creation.

While higher pay and support for skill requirements for these jobs may be initially logged as a higher cost, it can reduce costly and disruptive turnover. DSPs, for example, provide invaluable service for those with intellectual and developmental disabilities (IDDs), yet experience about 40 percent turnover rates within a year, as low wages lead them to be lured away by even routine private sector service jobs. Such disruption can set back progress in helping people with IDDs maintain jobs and can cause parents or siblings to have to temporarily drop out of the workforce to substitute for the help the DSP was providing. Those like the leaders of the Service Employees International Union and leading care and domestic worker advocate Ai-jen Poo have long argued that upgrading the pay, compensation, skills, and career paths of care jobs would not just reflect a proper valuation of the moral worth of such work but would also increase the productivity of such workers and bring higher returns to society overall.

Double Dignity Jobs Meet a Social Compact of Contribution and Work

The double dignity jobs approach is also far more in line with the basic American value of work than a UBI or UBC approach. A core philosophical aspect of President Franklin Roosevelt’s New Deal was that work would both provide purpose and dignity to workers in need of jobs and generate public support for an expanded social compact. FDR and his New Deal policymakers explicitly rejected unconditional support strategies, even when they were more efficient than jobs programs. They rejected unconditional payments as “repugnant to American ideals of self-reliance.” Labor Secretary Frances Perkins later wrote that FDR rejected hand-outs “when Americans wanted, above everything, to work and contribute.” As FDR described, work not only provides “a reasonable measure of security” for workers and for those who depend on them but it also has “moral and spiritual value.” This didn’t mean, as critics have claimed, paying people to dig ditches. Instead, it meant broader forms of work that contributed to society, from the Civilian Conservation Corps creating National Parks to “traveling libraries” for rural book delivery. While AI may force us to develop a new social compact, it should be rooted in the values of doing one’s part, contributing, and shouldering one’s share of the load. Call it a “compact of contribution.”

Many progressives resist the obsession with work-conditioned assistance for two legitimate reasons. One, so-called “work requirements” have been designed as a conservative strategy for denigrating Americans and reducing needed health care or food support due to excessively punitive paperwork burdens. Second, work requirements ignore the unpaid work of many Americans who care for a family member or raise children.

Progressives are right to fight tooth and nail against efforts to falsely stigmatize Americans and deny them needed benefits through punitive work requirements. But the answer to AI-led displacement of workers lies less in an unconditional cash-for-all approach like UBI and more in an expanded compact that honors American values of work, contribution, and carrying one’s share of the load. As Bharat Ramamurti notes, when asked about assistance in a world in which robots and AI can do almost all paid work, Americans opted two to one for aid being based on some form of effort or contribution as opposed to being unconditional.

A compact of contribution is the best path toward an AI era that properly recognizes the high value of currently unpaid care work and serious community service. There is undoubtedly a tension between the idea that caring for a child with a serious disability is considered “a job” when it is for a neighbor’s child, but not when done for one’s own child. The better path to moving toward an economy—especially in an AI era—that more properly recognizes the contribution of serious and often extraordinary unpaid care work is not calling for unconditional cash transfers. Instead, we should emphasize that such serious caregiving and exceptional community service honor a compact of contribution where all are shouldering their share of the burden. We should consider where it is appropriate to offer tax relief or even pay in some circumstances where there are clear standards for work, safety, and skills, and strong mechanisms to ensure program integrity and fraud prevention. The fully refundable child tax credit—that passed in 2021 as part of the American Rescue Plan—was one of the most successful economic dignity efforts ever because it both helped working class families overcome affordability squeezes and reduced child poverty to record lows. It must be reinstated. But even in making the case for this type of vital child allowance, there is likely to be more public support by stressing that more than 90 percent went to working families struggling with the cost of raising children and that virtually every parent or grandparent without income was either disabled or overwhelmed with significant care responsibilities. This expansion of our social compact based on individual responsibility and contributions beyond formal jobs was reflected in FDR’s Second Bill of Rights speech, when he called for “the right to command the necessities and amenities of life in exchange for work, ideas, thrift, and other socially valuable service.” (emphasis added). 

Expanding Double Dignity Jobs Requires Policy Flexibility, not a One-Size-Fits-All Approach

A dramatic increase in double dignity jobs must avoid a one-size-fits-all approach. In some cases, double dignity jobs creation can be fueled by policies like guaranteeing home care assistance for those with proven need, which would lower costs and elevate dignity for adult children and their older loved ones. This type of policy could be achieved with the necessary funding through Medicare or Medicaid to ensure there are sufficient well-paid care workers available to meet the expanded demand that would come with this new guarantee.

One could also imagine federal funding to states and cities to support high-value double dignity jobs in areas from navigators for those leaving the criminal justice system to college and career advisors to mental health experts or missing health professionals in poor or urban areas. With state and local flexibility would come maintenance of effort requirements to ensure that double dignity jobs funds are leading to net job creation and not fund diversion.

To be clear, a double dignity jobs approach should be seen as a macro-solution to expand meaningful, dignified work in an AI era and prevent weak labor markets from AI job replacement. This should not be seen, however, as a substitute for the policies described below in Part 3 that would help millions of displaced workers find their next jobs. This is not an unrealistic “coder-to-care worker” vision that pretends that every dislocated worker can or wants to go into a double dignity job. Nonetheless, some displaced workers may see opportunity in these jobs or will benefit from a spouse getting one. David Autor and co-authors have written that communities hit by the China shock eventually recovered employment levels through growth in health, education, and other service jobs, though this recovery took significant time and benefitted new workers entering the labor market more than the displaced manufacturing workers.

Beyond Double Dignity Jobs: We Should Incentivize the Private Sector to Favor People and Purpose

Beyond a dramatic commitment to create millions of new double dignity jobs, public policy should not be neutral or indifferent as to whether private sector companies choose job displacement over job creation or augmentation. One AI start-up filled billboards mocking employers that choose workers over AI, with banners like “stop hiring humans.”

A bus stop advertisement for the AI startup Artisan. (Artisan)

It has become standard fare for companies and management consultants to encourage companies to ‘think twice’ or pause before they hire or rehire due to AI replacement potential. For instance, Shopify’s CEO has initiated a policy that any employee seeking to expand headcount needs to prove why they “cannot get what they want done using AI” first. That may make sense for an individual company. And a democratic economy should not dictate private sector hiring or firing decisions. Yet, as a nation committed to the dignity of work, we should proudly promote a people and purpose preference. We should pass legislation like that co-sponsored by Senators Mark Warner and Josh Hawley to require quarterly company data on when and where AI displacement is taking place. Our economic policies should incentivize companies to choose augmentation over replacement, and to reinvest AI savings in new opportunities for high-value jobs at their workplaces.

We can start with the tax code. Daron Acemoglu, Andrea Manera, and Pascual Restrepo are among many who have found that our tax code incentivizes companies to favor technology over workers. They write that we effectively tax software and equipment at roughly 5 percent, while labor is taxed at 26 percent. If we are seeing a major job dislocation and reduction in net jobs, we should be aggressive in creating new tax policies that give tax advantages to those who choose job creation and the use of AI to augment rather than replace jobs.

Beyond the tax code, we should use public policy to encourage a people and purpose preference. In 2016, Brynjolfsson and McAfee founded the MIT Inclusive Innovation Challenge in part to encourage and recognize those who put people first in the face of AI and other technological disruption. One of the first winners was a health company that, instead of using new technology to simply replace workers doing routine tasks, used such IT systems to free up and empower “health coach” jobs to help guide patients on exercise, nutrition, and medical appointments. IKEA used AI to take over the traditional duties of call center workers, but then retrained many of those workers to become interior design advisers, helping drive even higher revenues for the company. When IBM built an internal AI tool to handle 94 percent of routine human resources tasks it reinvested the money into hiring even more workers for software engineering and sales and marketing. There can be opportunities for local and national governments as well. In Singapore, the government is creating plans to deploy driverless buses, while planning to use existing drivers and new workers “to help load luggage, aid elderly passengers in boarding, or lead a tour group” or work as “programmers, designers, fleet managers and operations control centre staff” that could enhance the quality of using buses.

Inspired by the MIT Inclusive Innovation Challenge, in 2020 I included a proposal in Economic Dignity for the creation of an ARPA-AI Jobs Initiative that would produce evidence and analysis on how AI can enhance jobs as opposed to replace them—with incentive grants to companies and local governments that work toward pro-job AI solutions. In 2022, Harvard Professor Dani Rodrik proposed a similar idea that would cover a broader set of employment-friendly technologies including manufacturing and energy jobs. An ARPA-AI Jobs Initiative should not wait.

Part 2: A Universal Economic Dignity Floor

Work is a necessary but not sufficient condition for a basic guarantee of economic dignity. Throughout U.S. history, low wages, lack of labor rights and protections to collective bargaining, excessive hours, and humiliating and unsafe work conditions have punctured holes in any number of economic dignity floors. Over the last 125 years, workers have won significant protections against domination and abuse at work. But devastating violations of economic dignity remain. Low wages still lead millions to work multiple jobs or excessive hours that force them to choose between putting food on the table and being present for the most meaningful moments of family life. Too many working parents still suffer the dignity harms of not being able to ensure decent housing or safe neighborhoods for their loved ones. As 50–70 percent of the changes in relative wages over the last few decades stem from automation displacing workers, AI could easily worsen wage inequality—especially if it contributes to a weak labor market with high unemployment. This is why we need not only more double dignity jobs and incentives for private sector companies to choose job augmentation over job replacement, but also an economic dignity floor for all who work and contribute.

We Can Have a True Economic Dignity Wage Without Hurting Jobs

A job’s capacity to provide the wage necessary for economic security is what Bayard Rustin described as “the pre-condition for pride and dignity.” The idea of work ensuring sufficient resources to provide for and care for family is hardly a far-left vision. In the Wealth of Nations, Adam Smith wrote about the need for wages to “bring up a family” in a way that is consistent with “common humanity.” Smith wrote that “they who feed, clothe, and lodge the whole body of people, should have such a share of the produce of their own labour as to be themselves tolerably well fed, clothed, and lodged.” And 113 years ago, President Teddy Roosevelt called for a “living wage” robust enough “to make morality possible, to provide for education and recreation, to care for immature members of the family, to maintain the family during periods of sickness, and to permit a reasonable savings for old age.”

There are several “pre-distributional” policies—measures that change the division of wealth and wages before revenues are collected or distributed—that are critical to securing economic dignity wages. These include non-compete clauses for workers, antitrust policies to prevent economic concentration that weakens wages as well as protections for small businesses and entrepreneurs from platform domination. Such policies work to help prevent the excessive concentration that could lead a fortunate few to extract unseemly amounts of AI-driven wealth that otherwise could flow to small business owners, medium-sized companies, and their workers.

While tools like stronger collective bargaining and bans on most non-compete clauses are essential, no pre-distributional policy has a more direct or widespread impact than a significant increase to the minimum wage. The consistent argument against a significantly higher minimum wage is that it would cause employers to reduce their headcount. But a large body of domestic and international research shows that recent minimum wage increases generally have little to no impact on employment. A true living wage significantly higher than $15 an hour will be a critical economic dignity floor for those workers in all industries and jobs that do not experience AI-induced productivity gains, or where they do but refuse to share equitably with workers.

Critics will argue that far higher wage floors in the AI era could further incentivize AI automation over human labor. While this concern has not yet materialized in past minimum wage increases—even after recent efforts around a $20 minimum wage—it is plausible that at living wage thresholds far higher than is imagined now there could be some job impacts. Yet this possibility should not be seen as a stop sign for an unprecedented minimum wage increase in an AI era. The possibility that a far higher minimum wage could cause marginally fewer jobs is not game, set, match against such an ambitious living wage. The economic dignity package being proposed here would include a significantly expanded national living wage together with a dramatic double dignity jobs initiative. Even if the Congressional Budget Office found marginal negative job impacts from an analysis of a higher living wage in isolation, it could still score a double dignity job and living wage package as raising wages and on net creating millions of new jobs.

An economic dignity wage floor does not have to depend solely on a minimum wage increase. Expanding the depth and breadth of the Earned Income Tax Credit (EITC) and reviving the expanded Child Tax Credit (CTC) would raise the economic dignity floor without inviting arguments that automation was being incentivized over human labor. Together, the combination of a higher minimum wage with an expanded EITC and CTC can create the right type of “guaranteed income.” Expanded tax credits or wage subsidies together with a true living wage can together ensure a guaranteed minimum income level for all workers, without sending everyone—regardless of need—the same equal payment under the guise of “guaranteed income” that would deplete resources needed for double dignity jobs creation, paid leave, refundable child tax credits and other building blocks of a universal economic dignity floor. 

Public Policies Could Facilitate a Shorter, More Productive Workweek and Work-Sharing if AI Reduces Labor Demand

John Maynard Keynes predicted in 1930 that we would hit a 15-hour work week by 2030. Keynes no doubt meant to be provocative. Yet he reflected the technology-optimistic belief throughout history that higher labor productivity can enhance our quality of life by allowing us to produce more in less time. The possibility of AI-driven productivity shortening work weeks without reducing pay should be seen as one of the most AI-optimistic scenarios for economic dignity. This AI-optimist scenario, however, comes with major question marks. In a world where there remains intensive corporate resistance to expanding collective bargaining, will there be appropriate expansion of collective bargaining or other pro-worker policies to ensure that workforce productivity gains are appropriately shared with workers and not just executives and shareholders?

If so, how would we balance the desire of workers who would prefer to work a single, 40-hour a work week job—with a higher paycheck due to shared productivity gains—with those who prefer the option to use higher wages through shared productivity gains to enjoy a four-day work week that allows more time for family and loved ones? These are issues that will have to be carefully weighed by worker advocates as they consider whether they should push a productivity trigger that would move the Fair Labor Standards Act to go to something like a 32-hour work week if an unprecedented productivity rise takes place over a set number of years.

To the degree that higher productivity is shared with workers but leads to some reduction in demand for human labor, we should consider a shorter work week (or “work-sharing”). Currently, our system of employer-provided health care may disincentivize that. Today, an employer who needs 4,000 hours of human work a week and employs 100 full-time workers may resist moving to 125 workers doing a 32-hour work week, because that employer would face a 25 percent increase in the workforce for which it must provide health insurance. If the level of disruption is significant enough that policymakers seek to encourage shorter work weeks, policymakers will need to either decouple health insurance from employment or provide tax incentives and subsidies that offset the costs of covering a larger workforce.

In the scenario of seeking to encourage more workers through a shorter work week, policymakers will need to consider how to move away from the current employer-provided model, or how to use tax incentives or subsidies to facilitate the option of a shorter work week. 

Promoting the Dignity of Work Means Not Denying Workers the Capacity to Experience the Most Equally Treasured Moments of Family Life

In a market economy with dramatic inequality, one of life’s equalizers is that our most treasured and meaningful experiences—welcoming a new child, caring for a parent in their final days—do not require riches or luxury. Yet a dramatic hole in our economic dignity floor is that a lack of resources—particularly paid leave—often denies tens of millions of Americans the capacity to enjoy these universally treasured experiences. Today, one in four employed mothers returns to work within three weeks of giving birth. While about four in five executives have paid leave for bereavement, only about one in five workers in the lowest wage occupations do. Within specific companies the leave allowed for the death of a child or spouse is defined not by a respect for universal human dignity but by the corporate job hierarchy. Paid leave is not just another program. It fills a hole in our economic dignity floor that denies tens of millions of Americans the most meaningful moments with family that could easily be guaranteed to all. As only those with lower incomes and employment benefits are denied the economic capacity to take such precious time, this is another area where targeted funds can create an economic dignity floor for all at a cost far below that necessary for universal payments.

Remember: It’s Economic Dignity Purchasing Power, Stupid!

In moving toward an economic dignity floor, it is critical to look not simply at inflation-adjusted wages—or even the capacity to buy regular goods like groceries and gas—but the degree to which a paycheck allows what most families consider the building blocks of economic dignity. Call it “economic dignity purchasing power” (EDPP). Today the strain that many working families feel is that even when their inflation-adjusted wages go up, the critical elements of providing a dignified economic life for their family—health care, childcare, and housing—are going up far faster. As KPMG reports, “between 1991 and 2024, the costs for daycare and preschool rose at nearly twice the pace of overall inflation.” From 2000 to 2020, rent rose faster than incomes in the counties home to 97 percent of Americans.

These price pressures are critical to defining an economic dignity wage. Policies that make health care, housing, and childcare more affordable increase EDPP. Consider childcare. For a family making $80,000 a year but paying $20,000 in childcare costs, a public policy limiting the costs of childcare to 7 percent of income—and thus saving them $14,000—raises their EDPP substantially. Increasing EDPP by reducing childcare costs is also pro-growth. Today, childcare is the most regressive tax on work and entrance into the labor force. Increasing EDPP for childcare could lead to a stronger labor force and higher family incomes. To the degree that AI makes the building blocks of economic dignity—like the costs of housing production—cheaper to build, our policies should ensure that AI-induced abundance results in lower costs that clearly raises EDPP for American families.

Beyond Wages, We Need to Expand Wealth Creation Opportunities for All Workers

AI has the potential to widen another major hole in our economic dignity floor: the extreme wealth inequality in our nation.

Today, the top 10 percent of American households hold two thirds of all wealth and 87 percent of all of the wealth in the stock market. The top 1 percent owns half of all stock wealth. Meanwhile, about 40 percent of Americans and well over half of Black and Hispanic Americans have no retirement savings plan. And while 81 percent of college grads have a retirement savings plan, just 39 percent of adults with no college education have one, according to a Gallup survey. Forty-seven percent of Black households have a net worth below $10,000, while 24 percent of Black households have a net worth of zero or negative.

Wealth inequality denies core aspects of economic dignity: a dignified retirement, the ability to age in place, a cushion or nest to help protect family members against economic disruption, and the potential to leave even a small something to children and grandchildren. AI threatens to make this worse. Many economists now forecast that if AI leads to increasing divergence of returns to capital over wages (especially if AI is used primarily for automation instead of augmenting human labor), the harm of this inequality in wealth and capital investment could grow.

Fortunately, much can be done to spread wealth creation more broadly without simply resorting to equal distribution regardless of need or current wealth as UBC proponents want. First, we know that our current tax system for retirement is upside down. By using tax deductibility as the main savings incentive, existing tax incentives give the most to those who need the least help saving and the least to those who need the most. A high-income saver gets to deduct 37 cents on the dollar, while lower income workers get nothing or only 10 cents. Second, those not considered employees or who do not work for employers who offer a 401(k) lose out not only on an attractive match, but the automatic, opt-out options that behavioral studies show create habits of savings. Georgetown researchers find that forty-seven percent of private sector workers “are estimated to lack access to retirement savings through their workplaces,” with lower-income workers much more likely to have no retirement savings plan through their job. And while three quarters of the top half of workers have access to a retirement saving plan through their job, close to two thirds of the bottom half of the income distribution do not.

It doesn’t have to be this way. Virtually all federal workers get these benefits through the Thrift Savings Program, with very positive results. The savings program reached 95 percent participation when it automatically enrolled workers, essentially eliminating the participation gap between high and low earners. Australia has gone national with a similar superannuation plan that J.P. Morgan reports has led that nation to have the fourth largest retirement system in the world, and projected to be second by 2031, while ranking only 55th in population.

As National Economic Council Director under President Clinton, I helped devise the USA Accounts proposal that he announced in his 1999 State of the Union, which sought to target precisely this hole in our wealth creation policies. The vision was to ensure that, on top of a rock-solid annuity through Social Security, all Americans could accumulate wealth through savings and the magic of accumulated returns by giving them access to the Thrift Saving Program that federal workers, including members of Congress and White House staff, participate in. To overcome the upside-down tax incentives, it offered those with very modest incomes $800 annually (worth $1,578 in November 2025 dollars), with matching that was more substantial for modest and low-income workers.

Progressives have tended to shy away from these universal wealth creation accounts for understandable reasons. First, while Clinton sought to define such accounts as universal pensions that are separate and additive to Social Security, many progressives fear that such proposals will be exploited by conservatives to privatize Social Security or used as a rationale for cutting Social Security’s guaranteed annuity benefits. Second, many progressives have felt it was tone deaf to tell hard-pressed Americans they should set aside a few thousand dollars for personal investment when they are choosing between food, prescription drugs, and fixing their car. It hardly helps families struggling with affordability to simply spend $2,000 less.

Yet these very legitimate fears should not prevent us from addressing the possibility that an AI future could further set back working families who do not gain the benefits of accumulated returns from capital investments. If so, progressive policy should focus on resisting the tax-deductibility incentive structure that makes up the contributions to the Trump baby-bond proposal. Instead, we should consider two-to-one matches for the lowest-income savers and ensure strong and novel efforts to increase the ease—including automatic, opt-out savings options—for those too often excluded, like care workers, household workers, and those in the gig economy.

AI Could Magnify Workplace Indignities Unless Policy Catches Up

Without the third leg of the economic dignity stool—freedom from domination and humiliation at work—economic dignity is unattainable. The desperation of working parents to put food on the table can force workers to subject themselves to workplace conditions that defy Kant’s fundamental definition of “dignity”: to be treated as an end in oneself, not purely as a means to someone else’s ends. Political philosopher Michael Sandel has described this as the “intrinsic value of being human.”

Historians point out that at the time of the founding, free citizens saw their threats to personal liberty mostly through the lens of their relationship to public authority—not to private-sector employers. Most free workers worked for themselves or in small shops. The American Industrial Revolution dramatically changed this. From 1800 to 1920, the percentage of people who were their own boss fell from 80 percent to 22 percent. With the rise of larger corporations, coal mines, and sweatshops, Americans found themselves asking a basic question: What good are individual political liberties if you are forced to endure what historian Elizabeth Anderson has described as a virtual dictatorship at work for the majority of your waking hours? The rise of the American labor movement reflected not just a fight for higher wages and benefits, but for the protection of a sphere of personal liberty and dignity in the workplace. The turn of Teddy Roosevelt from a more laissez-faire Republican to a champion of workers against corporate power emanates not from an abstract change in his political philosophy, but from his nighttime visits to worksites as New York City police chief. During these visits he was stung by the degradation and humiliation of workers and their children—whom he recognized as “the crushable elements at the base” of industrial America.

While techno-optimists see technological advancements as increasing human liberation, we have seen even recently the degree it can facilitate a hyper-efficiency of work through micro-management in ways that look more like domination. Some models that aim to extract maximum productivity from workers ignore basic elements of what it means to be human, like the need to rest, go to the bathroom, check on sick children, or experience moments of depression. In Economic Dignity, I cited accounts of meat-processing workers being forced to wear diapers because the relentless speed of the production line denied them bathroom breaks. Others reported urinating in garbage bags to avoid missing output requirements enforced by “electronic whips”—wristbands that track a worker’s every move and punish any deviation from excessive quotas and output requirements.

AI tools only expand the potential to track every minute or movement without, in the words of privacy and surveillance expert Chris Gilliard, taking “into account people’s dignity or humanity.” While AI tools can be used in reasonable ways to measure productivity and help workers work efficiently, extreme “digital monitoring” has resulted in workers facing consequences for any signs of inactivity over a 10-minute period, workers being “fired by an App” without explanation, and the use of AI tools to counter the right to collectively bargain.

While significantly expanding collective or sectoral bargaining to cover more workers would be the ideal way to negotiate which uses of AI are justifiable versus abusive, state legislation currently offers the most timely and realistic means to address threats to workplace dignity. This is evolving in real time. Proposed bills in California, Massachusetts, Vermont, and Washington would establish various safeguards against over-the-top surveillance and automated managerial decision-making. Bills in Colorado, Georgia, and Illinois would prohibit combining algorithms with workers’ personal data to set individualized wages.

This reality makes the Big Tech and Trump Administration efforts to impose a 5–10-year moratorium on state legislation one of the largest current threats to dignity at work. While national regulatory standards can at times have benefits for companies, workers and consumers, sound federal regimes typically develop only after states pioneer innovative approaches. In the case of AI, the only choice we currently have for worker protections is either a state-based “laboratories of democracy” approach or a national vacuum where pro-worker policies are blocked by a White House beholden to many of the largest AI corporate interests.

Part 3. A Just AI Transition Strategy for Workers

While there is significant disagreement among experts regarding the speed of AI adoption and resulting labor impacts, there is a widespread consensus that there will be significant labor disruption even if AI does not cause a net jobs loss. Policymakers should urgently develop policies to mitigate the negative impacts of these disruptions. Our failure to have significant policies in place to prevent or slow the speed of devastating downward spirals in communities facing factory closings caused severe harm to economic dignity over the past several decades. As Angus Deaton and Anne Case documented in Deaths of Despair, the greatest tragedy of globalization was not simply wage stagnation, but the catastrophic unraveling of the social fabric in manufacturing communities, leading to significantly increased mortality. That’s why if we are on the verge of the mother of all job dislocations, a comprehensive AI policy must support a dramatic expansion in policies that prevent deep dignity harms. We must ensure that the “deaths of despair” that plagued factory towns devastated by trade or technological change do not find a new echo in what could be a major white-collar job disruption due to AI. An AI transition policy should incorporate and expand the following lessons.

Ensuring That AI Promotes Basic American Values Should Not Be Seen as Slowing AI Down

Dramatic changes can be a lot less disruptive when communities have longer to adjust. Speed matters. Economist David Autor found that the communities most impacted by the China Shock were already economically vulnerable, but that the speed and intensity of the “shock” had devastating impacts. Indeed, one of the underappreciated policy failures of that era was that President Bush blocked the use of the “anti-surge” provision President Clinton secured in the U.S.-China Relations Act to pause job-destroying import surges from China.

Today, the push for a speed-at-all-costs AI policy has led major AI companies and the White House to dismiss anything that might serve as an “anti-surge” policy for AI job loss as undermining the United States in its race with China for AI dominance. Competing with China so that AI serves American values is a national imperative. We cannot allow the simple uttering of the words “race with China” to reflexively block any policy designed to shape AI to serve American values of work, freedom from excessive surveillance, and job protections. I used to say at the Treasury Department during the financial crisis that while fear of “systemic risk” was preeminent, simply uttering those words should not mean everyone’s IQ instantly drops by 50 points when judging any idea. Nvidia CEO Jensen Huang says in one breath that any state regulation will hurt our race with China, and with the next argues against bipartisan prohibitions on selling the most advanced AI chips to China even though such exports could be harmful to AI competition with China. State policies to provide more significant worker notice for AI-related job changes and to require things like discussions with worker representatives, information on AI-related job displacements, or taxes that encourage AI augmentation over AI job displacement should not be automatically rejected on the grounds that they will hurt the “race with China” any more than saying that states enforcing reasonable speed limits on AI driverless cars will hurt AI innovation.

Focus the AI Response on Individual Job Transitions and Individual Scarring

Unlike the factory closures that devastated entire manufacturing communities, AI displacement is less likely to cause the same kind of concentrated job loss and collapse of local economic demand that triggered downward spirals in many factory towns. Tech hubs like Boston or the Research Triangle may see higher job losses but will likely remain economically sound, benefiting from other AI developments.

While we need stronger place-based policies to address economic disruption in factory and energy communities—including expansions of Biden Administration innovations like Energy Communities, the Recompete Pilot Program, and the Domestic Conversion program—an AI transition policy will need to focus more on individual assistance to help geographically dispersed workers avoid devastating downward spirals.

A significant body of research demonstrates that long-term unemployment creates deeper health and emotional scarring than almost any other events in life other than the death of a child or spouse. Surveys further document the degree to which those who struggle with long-term unemployment often face discrimination in finding new work as their length of unemployment and worse credit scores screen them out of getting job interviews, leading to depression and declines in self-esteem. In addition to the expansion of job and career assistance policies discussed below, we should put forward special policies that include prohibitions on hiring algorithms that discriminate based on length of unemployment or indications of long-term unemployment like credit reports, and we should pass legislation like Senator Chris Van Hollen’s Long-Term Unemployment Elimination Act to subsidize job opportunities and wrap-around services for those out of work for extended periods.

AI Should Drive Us Toward a Universal Workforce Policy That Doesn’t Rely on How You Lost Your Job

The paradox of AI-related job disruption is that it both cries out for an immediate, AI-specific solution while also demonstrating why our nation needs universal job transition strategies that do not ask about the cause of a job-seeker’s career disruption.

While there will no doubt be an immediate need for policies targeted toward those who lose jobs due to AI, in the medium-long term, we must recognize that such policies will continue an unfortunate legacy of linking the degree of adjustment assistance to what caused job loss instead of simply asking how to help workers find a new one.

The only federal program that, until recently, provided significant enough income support to help workers train for new careers was Trade Adjustment Assistance (TAA). Yet, it had a critical design flaw. Workers who lost jobs could get income assistance only if they could prove that their job loss was caused by trade. While helping workers displaced by trade is essential, limiting serious retraining assistance to only those who can prove the cause of their job loss was trade-related hurt its effectiveness and created enormous horizontal inequity. If three workers all lose their jobs and careers—one through automation, one through bankruptcy, and one through trade—why should only the third get significant assistance? This focus on the cause of job loss also leads to critical operational deficiencies. Former Biden and Obama administration official Jacob Leibenluft in a forthcoming Washington Center for Equitable Growth paper on the history of TAA, notes that such nuanced inquiries into whether certain workers’ job loss was truly trade-related, often delayed immediate assistance when it was most needed.

AI can compound these inequities if new assistance is limited only to those whose job loss can be proven to be AI-caused. For example, it could be seen as inequitable if AI vulnerable workers like software developers who typically make over $130,000 and personal financial advisors who typically make over $100,000 get significant adjustment assistance or wage insurance while such aid is denied to workers who made $50,000 but lost their jobs for non-AI reasons.

Another flaw in linking assistance to those whose job is displaced by AI is that it fails to address many of those who feel most economically threatened by AI—those who have chosen careers in job categories that AI now threatens to eliminate, such as entry-level, white-collar jobs. Call them “the pre-displaced.” In August, Stanford’s Digital Economy Lab published findings showing that employment for those ages 22–25 in occupations most susceptible to AI exposure has experienced a 16 percent relative decline. While some have disputed these findings, what is generally recognized is that AI is likely to disrupt current career ladders, where entry-level white-collar jobs require rookie workers to engage in “tedious work in exchange for valuable experience.” Students who have devoted years to training for a specific career may soon enter the labor force only to discover that their training has rapidly been rendered obsolete. Further, we know that young adults whose first jobs are delayed by a weak job market often face long-term economic losses—or “scarring.”

All these equity and definitional issues point in the same direction: National adjustment policies should focus on how to find new careers and jobs, as opposed to why a job was lost.

Elements of a Universal Worker Adjustment Policy Framework:

• Make the Benefits of Trade Adjustment Assistance. When it was fully operational, TAA covered a very small share of displaced workers. Many, of course, found jobs quickly on their own or benefitted from cost-effective job search assistance. An expansion of wage insurance beyond the Obama-created assistance for those over 50 years old will do more to incentivize workers to seek immediate new jobs. Making this program available to all—regardless of the cause of job dislocation—will be a major benefit to the AI-displaced without causing delays or horizontal inequity for those who lose their jobs for other reasons.
• Close Trust Gaps with Job-at-End-of-Road Training. Engaging with workers and employers over a few tours in the White House, I have been struck by the mutual distrust of training programs. Private-sector companies doubt that government programs are sufficiently employer-driven to impart the skills necessary to contribute to their businesses. Workers, on the other hand, are reluctant to sacrifice their time and spend precious dollars on childcare when they distrust that a good job will be waiting for them at the end of the program. Sketchy for-profit colleges have only added to this distrust. A more universal expansion of job and career adjustment should therefore be oriented towards job-at-the-end-of-the-road policies. This can start with major expansions of apprenticeship and pre-apprenticeship plans in broader fields including, but also in addition to, successful Building Trade programs. Funding should be significantly increased for sectoral-training partnerships like Project Quest and the Wisconsin Regional Training Partnership—both of which have proven track records in occupational training for clear local needs—and employer-community college partnerships for needed skills in open jobs. While many tout the need for portable industry credentials, actual progress in establishing nationwide standards has been relatively slow. Entrepreneurial assistance programs that help displaced workers move toward self-employment can also alleviate the economic dignity harms caused by transformative AI. While there is understandable fear of programs that subsidize on-the-job training for expenditures corporations would have borne anyway, there should be an increased openness to helping to fund on-the-job training or programs where workers work part-time while also being funded for their credential- and skill-building programs.

• An Increased National Commitment to Non-College Degree Career Building. Recent years have seen a growing bipartisan consensus for a dramatic expansion of federal support for career building outside of the four-year college degree. These efforts include proposals for free community college, short-term Pell Grant support for targeted skill-building, and a greater emphasis on industry-wide credentials. And they include job-at-the-end-of-the-road initiatives discussed above and policies championed by the Opportunity Network to dismantle “paper ceilings” where workers have gained necessary skills through routes outside of traditional college degrees. There is growing recognition of the success and popularity of initiatives like IBM’s P-TECH, which partners with public schools and community colleges to provide students with mentors and paid internships, ensuring a direct pipeline to high-growth industry jobs. Trade schools in Massachusetts are facing unprecedented waiting lists. The beauty of such expansions is that they provide not just universal support for those displaced, but the pre-displaced and those who have been out of the workforce or who seek to push their potential by pursuing better, high paid work or ones that provide greater purpose and meaning at work. With increased funds from AI productivity growth, this is another case where targeted support for career-building does far more to create an economic dignity floor than universal payments spread out to all.

AI Companies Should Fund AI Specific Job Adjustment Policies

As much as AI job disruption should be an impetus for a universally available job adjustment and career education initiative, this is no time to make the perfect the enemy of the good. A major first step would be a five-year initiative focused on AI job displacement funded by a tax on AI developers and employers reducing jobs while benefiting from AI. For instance, Senator Mark Kelly’s proposed AI Horizon Fund offers one model—a federal trust financed by AI industry revenues to invest in worker training and transition support.

An AI tax that focuses less on AI users and more on those profiting from AI development or making employment choices can provide a dedicated stream for funding assistance for those who are AI displaced. To the degree that an adjustment strategy is to be limited only to those impacted by AI—and not job loss or unemployment or underemployment due to other causes—it would make more sense to use funds from the companies that are benefiting from AI development and those choosing to use AI for job replacement as opposed to job augmentation.

There is no shortage of AI leaders expressing confidence that AI will broadly lift human well-being through medical breakthroughs and higher productivity. There is also no shortage of analyses on the potential degree and speed of AI’s impact on job tasks, job categories, and ultimately, the nature and number of jobs themselves. What there is, however, is a shortage of the most influential private sector leaders and Administration officials treating AI’s impact on the livelihood, economic security, sense of purpose, and social connection of millions of workers as the preeminent AI issue. The impact of AI is too often treated as an issue to be observed, not shaped.

If we the people—not tech giants—are to shape our economic destiny, we need to start with clarity on our aspirations. Our North Star must be a commitment to economic dignity rooted in a path for all Americans to pursue purpose, potential, meaning, and economic security through work and contribution to family, community, and country. We will need to build support to share the productivity gains from AI and to use those resources wisely—not simply deplete them without targeting help to those facing barriers and to the challenges that today deny dignity to so many. This means creating millions of jobs with dignity that give other Americans needed dignity and opportunity. It means plugging policy holes that prevent all Americans from being able to count on a basic economic dignity floor. And it means leaving no segment of Americans behind when we face massive economic and technological disruption.

Uncertainty abounds—but we don’t need to wait. With clarity on our North Star, we can further policies that move us in the right direction, modulate them as we learn more about the speed and severity of AI’s job impacts, and help every one of us to experience the economic dignity that all humans deserve.