The debate about AI and jobs tends to focus on one question: how many jobs will it eliminate? The quantity of jobs is undeniably important, but John Chachas is more concerned about the quality of jobs. What meaningful work will remain as the supremacy of artificial intelligence grows?
Chachas is the founder of Methuselah Advisors, a boutique investment banking firm that has spent decades advising major media and technology companies through acquisitions, restructurings, and the corporate transformations that precede them. He has watched, from inside boardrooms, how companies shed labor when the economics shift. And what he sees in the current labor market is not a crisis of scarcity. It is a crisis of character.
"We are heading toward many fewer real jobs and many more make-work jobs," Chachas has said. "That is not a very satisfying future."
Make-work. The phrase is worth sitting with. Jobs that exist not because they represent genuine economic value, but because they haven't been automated yet. Jobs that don't teach a craft, don't lead anywhere, and don't pay enough to justify the cost of the degree that often preceded them. By Chachas's reckoning, those are the jobs being created now, and AI is accelerating the process of replacing the real ones.
American Workers Feel Less Engaged in Their Jobs
Recent surveys of American workers suggest that the make-work phenomenon may already be influencing job satisfaction.
According to Gallup's 2026 State of the Global Workplace report, just 31 percent of workers in the United States and Canada report feeling genuinely engaged with their role, down from a post-pandemic peak of 33 percent in 2021 and 2022. Globally, engagement dropped to 20 percent in 2025, falling from a 2022 peak of 23 percent, the second consecutive annual decline and the first time it has fallen for two straight years. The $10 trillion in lost productivity Gallup estimates this costs the global economy annually is one measure of what it means. A workforce going through the motions is another.
The Conference Board offers a complementary picture. Workers are less satisfied with nearly every aspect of their employment than they were a year ago. Pay and promotion are the most barren ground: only 30 percent of workers report being highly satisfied with their compensation, and roughly 25 percent are satisfied with their prospects for advancement.
The Gallup data also shows a particular toll on managers, whose engagement has dropped nine points since 2022, erasing what had historically been a meaningful engagement premium over the workers they lead. These are the people responsible for developing and retaining the entry-level talent coming in beneath them. When they disengage, the institutional investment in younger workers tends to follow.
Taken together, these numbers describe something more specific than general dissatisfaction. Workers are not simply unhappy with their pay or their managers. They are disengaged from the work itself. What the data captures is a workforce that has stopped believing their jobs lead somewhere.
People can tolerate difficult work, low pay, and demanding bosses when they see a future in what they are doing. What they cannot sustain, at least not without disengaging, is the sense that the work they are performing every day does not really matter.
The AI Effect: Labor Optimism Declined After Chat GPT’s Release
ChatGPT launched in late November 2022. The before-and-after picture of the labor market is instructive, not because AI is necessarily the cause of everything that followed, but because it clarifies the trajectory Chachas has been describing.
In mid-2022, 71 percent of North American workers told Gallup it was a good time to find a quality job. Job openings were near a historic peak of 12 million. The quit rate was elevated, signaling that workers felt confident enough to leave for something better.
Today, only 47 percent of workers in the United States and Canada say it is a good time to find a job, a 23-point drop from that 2022 peak, placing the region second-to-last in Gallup's global rankings. Job openings have fallen from that 12 million peak to 7.1 million as of late 2025. According to job search platform Adzuna, entry-level vacancies, covering graduate roles, apprenticeships, and junior positions, have fallen 32 percent since ChatGPT's launch. Their share of the total job market has dropped from 28.9 percent in 2022 to just 25 percent. Employment in AI-exposed occupations for early-career workers has declined 16 percent over the same period. Software developers specifically have seen employment fall nearly 20 percent.
The unemployment rate for recent college graduates has gone from roughly 3.9 percent in 2022 to 5.7 percent in the fourth quarter of 2025, nearly a 50 percent increase, in a period when the overall unemployment rate barely moved.
Chachas's insight here is not just the data. It is what the data describes. Companies replacing entry-level analysts with AI tools do not issue a press release. They simply stop posting the role. The headcount shrinks quietly. The job that was the first rung on the career ladder disappears, and the rung above it, with no one to climb to it, loses its purpose too.
"Anyone arguing that AI is going to open a world of creative genius that presents more options for people," Chachas has said, "I don't see that."
John Chachas’ Solutions to the “Make-Work” Economy
Start With Education
Chachas believes the solution to the make-work problem can be found at the source: universities. In the past, earning a four-year degree was seen as a ticket to a higher-paying career. According to Chachas, higher education credentials are worth less now, despite costing more.
“If I had to do it over again today, I would be very hard pressed to say that spending $75,000 a year to get your undergraduate degree at Columbia and then another $85,000 a year for two years to get your MBA someplace means you can snap your fingers and it's suddenly going to point you into an employable direction in the American economy,” Chachas said.
Chachas's concern is not with learning broadly. It is with the mismatch between what many programs teach and what genuinely meaningful, AI-resistant work actually requires. The tasks most vulnerable to automation are precisely those that define entry-level professional work as it currently exists: drafting routine documents, processing structured data, writing first-draft code, running templated analyses. These are the tasks most four-year programs implicitly train students to do. Preparing young people for that pipeline, in a world where AI is rapidly absorbing it, is not a sound investment.
Chachas believes the answer lies less in traditional credentials and more in specific, demonstrated capability.
"I think we're in an era where stacked certifications in a specific, narrow category are going to be far more interesting to the employer than you graduated from Georgetown with your degree in political science," he has said.
The value of a degree, he argues, has been quietly decoupled from the value of the skills it is supposed to represent. Employers are increasingly looking for the latter and have growing tools to find it without the former.
The more durable preparation is also the harder one. Critical thinking, ethical reasoning, the ability to navigate ambiguity, relational intelligence, and the capacity to lead and persuade real people through complex problems are what AI consistently fails to replicate. They are also the skills most likely to make a career genuinely irreplaceable rather than merely unfired.
Chachas thinks universities owe their students a direct accounting of this reality. "Colleges owe it to their students to ask the question: 'What do you think you're going to get out of this? Is this worth it financially to you and your family to get this credential? And how will it employ you and advance your life?'"
Getting away from the assumption that a four-year degree is the automatic answer, he argues, is the starting point for building an education system that actually prepares young people for the economy being built around them, rather than the one that no longer exists.
Universal Basic Income
Chachas's analysis leads to a specific policy position, one that separates him from most commentators in this space. If corporations are going to capture the productivity gains that AI delivers, he argues, they should bear a proportionate share of the social cost those gains create.
His proposal is a compulsory UBI trust fund. Companies that deploy AI in ways that demonstrably eliminate human employment would be automatically liable for contributions into a federally managed fund. The more jobs a company removes through automation, the more it pays in. The fund would provide a meaningful income floor for displaced workers.
The economic logic matters as much as the mechanism. Workforce reductions that transfer productivity gains entirely to shareholders while pushing the cost of unemployment onto public programs are, in practical terms, a subsidy. Taxpayers fund the social infrastructure that absorbs the workers a corporation has replaced. The trust fund would make that relationship explicit and redirect the benefit toward the people who actually bore the cost.
Crucially, Chachas frames this not as a guaranteed income for inactivity but as a floor that makes meaningful work possible. The reason make-work persists is partly economic: people take hollow jobs because the alternative is nothing. A UBI that meets basic needs gives workers the stability to retrain, to start something, to find or create work that matters rather than work that merely pays. The goal is not to replace the labor market. It is to ensure that the people the labor market displaces are not simply abandoned to it.
The policy conversation in Washington has largely sidestepped this kind of accountability for AI adoption. The dominant proposals center on retraining programs and educational initiatives, which place the entire burden of adjustment on workers rather than on the companies that profited from the automation. Chachas thinks that asymmetry is both economically inaccurate and politically unsustainable.
"Has anyone on the Hill started to talk about this concept of UBI and how we could fund it?" he argues. In his view, “if your corporation deploys AI that destroys human employment, you will automatically be liable for payment into a UBI Trust Fund.”
The make-work economy is not inevitable. It is a policy choice dressed up as economic reality. The jobs being created now are, in many cases, jobs that exist only because there is no floor beneath them. Build the floor, and the incentive structure shifts. Companies that automate and pocket the gains would be required to fund the transition they created. Workers freed from hollow employment would have the stability to find or build something real. The distinction between those two kinds of work, Chachas has spent years arguing, is the distinction that matters most.
