AI, gig work, basic income and a banking system built around the traditional paycheck are beginning to collide
From the Craig Bushon Show Media Team
For generations, Americans have been conditioned to look at a single number when they want to know whether the labor market is healthy. If unemployment is low, we assume people are working, and if it rises sharply, we assume something has gone wrong. That relationship is becoming considerably more complicated than the headline suggests.
In June 2026, the official U.S. unemployment rate declined from 4.3% to 4.2%. Taken by itself, that sounds encouraging, and plenty of coverage treated it that way. Look underneath the number and the picture changes substantially. According to the Bureau of Labor Statistics household survey, employment fell by 507,000 people during June while the civilian labor force declined by 720,000. The number of Americans classified as “not in the labor force” increased by 832,000 in a single month. Labor-force participation dropped three-tenths of a point to 61.5%, the lowest reading since March 2021 and, excluding the pandemic years, the lowest since 1976. The employment-to-population ratio slipped to 59.0%.
That produces a situation which sounds almost contradictory, because employment can fall at the same time unemployment falls. The reason is straightforward once you understand what the statistic actually measures. The unemployment rate does not count everyone without a job; it counts people without jobs who remain in the labor force and are actively looking for work. When someone stops looking, that person moves out of the calculation entirely, and the rate can improve for reasons that have nothing to do with improving conditions. That distinction is not academic. Alongside the 7.1 million people officially counted as unemployed in June, roughly 6 million Americans classified as outside the labor force told the BLS they currently wanted a job, and not one of them appeared anywhere in the 4.2% headline rate.
None of that is automatically alarming. Millions of Americans outside the labor force are retirees, students, caregivers, people with disabilities, and people who simply don’t need or want employment. The usual explanations for a falling participation rate are an aging population and a shrinking immigrant workforce, both of which are real and both of which have been pushing the number down for years. What makes June worth a second look is that the steepest decline came from prime-age workers between 25 and 54, whose participation rate fell six-tenths of a point to 83.3%, with adults between 25 and 34 leading the drop. Retirement does not explain a 28-year-old leaving the workforce. Some economists have cautioned that a single month can be statistical noise that reverses, and that caution is warranted. But it raises a question that is going to get louder rather than quieter: what happens if technological displacement causes more working-age Americans to stop participating in the traditional labor market altogether?
AI Doesn’t Have to Produce 15% Unemployment to Transform the Labor Market
There is a tendency to imagine technological unemployment as another Great Depression, with millions of people suddenly out of work and the unemployment rate exploding across a single grim quarter. That is probably not how this happens.
The International Monetary Fund has estimated that roughly 40% of jobs globally are exposed to AI-driven change, with exposure reaching approximately 60% in advanced economies. It matters that “exposed” does not mean eliminated. Some workers will become dramatically more productive because of these tools, many jobs will change in character without disappearing, and some tasks and positions will vanish outright. The transition will therefore arrive gradually and unevenly, through thousands of ordinary business decisions that individually make no news at all.
A company employing 1,000 people discovers it can accomplish the same work with 850. Another business fires nobody but declines to replace the fifty employees who retire or leave. A dealership finds that AI can identify customers, communicate with prospects, schedule appointments, value vehicles and handle administrative work that previously required several people on payroll. A corporation assigns AI agents the analysis it used to hand to entry-level hires. A warehouse adds robotics, and a call center needs fewer representatives because software handles the routine calls. No single one of those announcements transforms the American labor market, but multiply them across millions of businesses over five or ten years and something significant happens: the economy keeps growing while requiring fewer human labor hours for every dollar of output.
That would be enormously beneficial for productivity, and it would create a problem that receives far less attention than the productivity gains do. If the economy increasingly doesn’t need someone’s labor, how does that person get money?
The Question Nobody Can Avoid: How Are People Paying Their Bills?
People do not stop needing housing because they are no longer counted as unemployed. They still need food, electricity, transportation and healthcare, and if somebody isn’t working a conventional job, the money has to come from somewhere. Today that generally means a spouse’s paycheck, Social Security, retirement savings, disability benefits, government assistance, family support, investment income, accumulated savings, debt, or some combination of all of them.
Increasingly, another answer is emerging, and it is worth understanding on its own terms rather than dismissing it as a side hustle. Gig and variable income are becoming significant enough that America’s financial system can no longer treat them as fringe sources of earnings, and gig work itself is evolving well beyond what most people picture. Americans hear “gig economy” and think about driving for a rideshare company or delivering food, but the more consequential development is that platforms have begun fractionalizing conventional employment itself.
Consider Shiftsmart, whose platform allows independent contractors to select individual work shifts through an app. The company’s materials describe store-associate work, retail auditing, merchandising, food preparation, restocking, inspections and mystery shopping, with workers completing onboarding, choosing available shifts and generally getting paid within days. Shiftsmart describes its own business with a line that deserves attention. “The Future Of Work Is Shifts, Not Jobs,” the company says, and it pitches its platform to employers as a way to fractionalize jobs into shifts.
Think carefully about what that model implies at scale. Instead of one employer for forty hours every week, a person constructs an income out of pieces: three hours on a retail shift, four hours of delivery work, five hours on another assignment, a freelance project, some online work, maybe a seasonal position when the calendar allows. The work hasn’t disappeared, but the job as a unit has been taken apart, and that distinction is going to matter enormously for everything built on top of the old unit.
Then Comes Universal Basic Income
This is where the conversation becomes politically uncomfortable, though avoiding it does not make the underlying economics go away. If AI and robotics eventually allow the economy to produce substantially more with substantially fewer workers, the United States will have to confront a distribution problem, because capitalism does not merely require production. It requires customers.
A robotic factory can manufacture an extraordinary volume of goods with a skeleton crew, AI can strip out much of the cost of administering a company, autonomous systems may eventually cut transportation costs, and artificial intelligence can already perform portions of professional work at very low marginal cost. That all sounds economically fantastic until you ask who is buying. The machines don’t buy houses, the AI doesn’t purchase pickup trucks, and the robot doesn’t take its family out to dinner on Friday night. Consumers do that, and for most Americans employment has been the mechanism connecting production to consumption. You provide labor, you receive wages, and you spend those wages on what other people and businesses produce.
If technology substantially weakens the link between labor and production, we eventually have to answer how purchasing power gets distributed in an economy that doesn’t require as much human labor. Universal basic income is one possible answer, and it is not the only one. There are enormous unresolved questions about affordability, taxation, work incentives, inflation and the appropriate size of any such program, and anyone who waves those away isn’t being serious. But dismissing the idea without addressing the underlying problem misses the point entirely. If AI generates large productivity gains while displacing enough labor income, some mechanism will face growing political pressure to replace part of what was lost, whether that turns out to be a basic income, a negative income tax, wage subsidies, expanded tax credits, social dividends, broader ownership of productive capital, shorter workweeks, or some policy nobody has drafted yet.
Now imagine what happens to our statistics if any version of guaranteed income becomes normal. A person receives a basic income, works ten or fifteen hours a week through gig platforms, and picks up occasional freelance income. That person may have enough money to live on, may not want a conventional forty-hour job, and may not be looking for one. Under our current framework, that individual’s economic life looks nothing like the twentieth-century employee around whom most of our statistical and financial infrastructure was designed.
Then the Banks Have a Problem
America’s lending system was built around predictable employment. When someone applies for a mortgage or an auto loan, the lender traditionally wants to know where the applicant works, how long they’ve been there, what they earn, what they owe and how they’ve handled credit in the past. Every one of those questions assumes a job sits at the center of the borrower’s financial life.
Consider a future borrower receiving $1,500 from a hypothetical basic-income program, $1,800 from gig work, $900 from freelance projects and $1,200 from part-time employment. That is $5,400 in monthly cash flow, roughly $64,800 a year, with no conventional job anywhere behind it. Under today’s underwriting conventions, that borrower can be considerably harder to evaluate than someone earning the same $64,800 from a conventional salary, because each income stream may carry different documentation, history and stability requirements.
Mortgage lenders are already running into pieces of this problem. Fannie Mae surveyed nearly 200 senior mortgage executives in late 2024 about variable and digital gig income, and the results describe an industry caught between recognition and capability. Sixty-seven percent of those lenders said accepting these income types would improve consumers’ access to credit, and nearly half reported growth in the number of borrowers using them to qualify. At the same time, 83% said digital gig income was difficult to use in approving applications, citing income history and stability requirements, calculation and documentation problems, and the absence of detailed industry standards. The workforce, in other words, is evolving faster than the financial infrastructure built around it.
Technology may eventually help solve the problem technology helped create. Federal financial regulators have already recognized bank-account cash-flow information as a legitimate form of alternative underwriting data, which allows lenders to evaluate how money actually moves through someone’s accounts rather than relying exclusively on conventional credit files. That could change the fundamental lending question from “what is your job?” to “how reliable is your cash flow?” An AI underwriting system could examine years of transaction history, income volatility, savings, reserves, recurring obligations and repayment behavior, and reach conclusions that today’s process cannot. Someone without a conventional job might turn out to be an excellent borrower, while someone earning $100,000 from a traditional employer, spending every dollar of it and carrying heavy debt might represent considerably greater risk. Creditworthiness and employment status begin separating from one another, and once they do, the entire consumer credit system has to be rethought.
We May Need a New Economic Dashboard
All of this is why we should be careful about looking at a 4.2% unemployment rate and concluding that the American labor market is healthy. The unemployment rate remains important, but it may no longer be sufficient on its own.
If the economy is moving toward some combination of artificial intelligence, automation, fractional work, gig income and eventually some form of income support, we will need a broader dashboard for measuring household economic health. That dashboard should tell us what percentage of working-age Americans are actually employed and how many hours they are working. It should break household income into its sources, showing how much comes from wages, how much from gig platforms, how much from government transfers, and how much from investment and capital. It should track how much household consumption is being financed with debt, and how many people want more work but have given up looking for it. Most importantly, it should answer whether the typical American household can maintain its standard of living on earned income. That last question may eventually tell us far more than the unemployment rate ever did.
The Transformation Is Bigger Than Jobs
There is a tendency to reduce the entire AI debate to one question about whether the technology will take your job. That framing is becoming far too narrow for what is actually underway. AI is changing what a job is, gig platforms are changing how work gets distributed, automation is changing how many people businesses need, income-support mechanisms could change the relationship between employment and survival, and AI-driven financial systems could change how banks decide whether someone qualifies to buy a house, finance a vehicle or start a business. Those are not five unrelated stories. They are different parts of the same economic transition.
The industrial economy gave us factories and mass employment. The information economy gave us computers and knowledge work. The AI economy may give us extraordinary productive capacity while forcing us to reconsider a system in which employment has been the primary mechanism for distributing income. Nobody knows exactly where that transition ends, and anyone claiming certainty about how many jobs AI will ultimately create or eliminate is getting well ahead of the evidence.
Reading Between the Lines
We don’t have to know the final destination to recognize the direction of travel. The most important number of the next decade may not be the unemployment rate at all. It may be the percentage of Americans who can still generate enough reliable income from their own labor to support themselves and their families. If that number starts falling while productivity keeps rising, this country will be facing a far bigger question than whether AI is taking jobs. We will have to decide how an economy built around work functions when work itself is no longer required in the same quantities, and that is a conversation worth having now rather than after the fact.
The Craig Bushon Show
We don’t just follow the headlines… we read between the lines to get to the bottom line of what’s really going on.
Disclaimer: This commentary is presented for educational, informational and editorial purposes. Discussion of potential future developments involving artificial intelligence, employment, universal basic income, lending practices and government policy represents analysis and informed speculation rather than predictions of certain outcomes. References to companies or platforms are illustrative and do not constitute endorsement. Financial and lending decisions should be evaluated based on applicable laws, regulations, underwriting standards and individual circumstances.
