The claim is striking: 35% of American workers are “disposable.” It’s also exactly the kind of statistic that demands scrutiny before we start treating it as a diagnosis of the American economy.
Paul Osterman’s argument is straightforward. Companies increasingly want the labor without the obligations that traditionally came with employing people. Contractors, freelancers, temporary workers, gig workers, and low-investment employees give companies flexibility while shifting risk onto workers.
That trend is real. But the article’s framing risks turning a complicated labor-market transformation into a morality play: corporations are the sharks, workers are the prey, and regulation is the barrier.
The reality is messier.
1. “Disposable” is doing a lot of work
The biggest problem is the category itself.
Osterman groups contractors, freelancers, gig workers, temporary employees, and “marginal” employees into one enormous bucket. But those groups have radically different experiences.
A freelance designer who deliberately chooses project-based work is not equivalent to a warehouse worker whose hours are manipulated to prevent full-time status. A highly paid contract attorney is not equivalent to a delivery driver barely clearing minimum wage.
Calling all of them “disposable” may make for a powerful headline, but it can obscure the distinction between flexibility chosen by workers and insecurity imposed on workers.
That distinction matters enormously.
2. The 35% figure needs more interrogation
The article’s headline statistic comes from Osterman’s own survey of more than 6,000 workers. That is potentially valuable evidence—but the definition of “disposable worker” determines the result.
If part-time workers are included because employers deliberately create high-turnover jobs, for example, the number suddenly becomes much larger than what most people would intuitively understand as “disposable labor.”
That doesn't make the statistic wrong. It means readers need to know exactly what they're being asked to believe.
A statistic can be technically defensible and still rhetorically misleading.
3. Flexibility isn't automatically exploitation
The article acknowledges that freelancers report higher satisfaction than standard employees. That is a rather inconvenient fact for the central thesis—and it deserves more attention.
People value autonomy.
Some workers prefer choosing their clients, controlling their schedules, working remotely, or moving between projects rather than climbing a corporate career ladder. For them, the absence of a permanent employer isn't necessarily a failure of the labor market.
The real problem isn't nontraditional employment.
It is nontraditional employment combined with inadequate bargaining power, unstable income, poor working conditions, and weak protections.
Those are not the same thing.
4. The productivity argument is more interesting than the moral argument
The strongest case against excessive disposability isn't that corporations are behaving badly.
It's that disposable labor can make organizations worse.
If workers are constantly cycling through an organization, they accumulate less institutional knowledge. They have weaker relationships with colleagues. They may be less invested in safety procedures. Managers spend more time recruiting and training replacements.
The article cites research linking contractor-heavy hospital cleaning arrangements with higher infection rates and disposable workforces with higher injury rates.
If those relationships hold up broadly, the economic argument becomes much stronger:
Companies may save money on labor while quietly increasing the costs of turnover, mistakes, accidents, and poor coordination.
That's a more compelling critique than simply saying corporations don't respect workers.
5. AI may accelerate the trend—but it didn't create it
This is one of the article's better points.
The disposable workforce didn't suddenly appear because of ChatGPT. Companies have been outsourcing, contracting, automating, and restructuring employment for decades.
AI may simply make the incentive stronger.
If management genuinely doesn't know what jobs will exist in two years, it has an obvious reason to avoid making long-term commitments today. Why build a permanent workforce around skills that might become obsolete?
That doesn't necessarily mean AI will destroy employment.
It may mean companies increasingly want optionality.
And workers, unfortunately, don't get to live on optionality. They need rent money next month.
6. “Just unionize” isn't much of a strategy
The article eventually arrives at unions, worker organizing, public pressure, and clearer employment rules.
Those are reasonable tools. But the prescription is considerably less developed than the diagnosis.
Unionization has structural limitations in an economy where employment relationships are increasingly fragmented. A contractor may technically work for one company, legally belong to another, and perform labor inside a third organization's workplace.
Who exactly is the employer?
Who negotiates?
Who pays?
Who is responsible when something goes wrong?
Those questions are becoming harder precisely because companies have become extremely good at separating control of labor from responsibility for labor.
That is the real policy problem.
The bigger issue isn't disposable workers. It's disposable responsibility.
The most useful takeaway isn't that America has somehow created a third of a workforce that companies can casually throw away.
It's that the American economy has become increasingly skilled at externalizing the risks of employment.
Companies want flexibility.
Workers want income and security.
Consumers want low prices.
Investors want returns.
And governments are left trying to decide who should absorb the cost when those interests collide.
That's where the argument should go next.
The question isn't whether disposable work should exist. It obviously should—for workers who genuinely want it and businesses that genuinely need it.
The question is much harder:
How much flexibility should companies be allowed to buy by transferring risk onto everyone else?
If the answer is “as much as the market will tolerate,” then 35% may not be the ceiling.
It may be the beginning.