MIT Calculates 50 Million Disposable Workers in the American Economy
More than 50 million Americans — or 35% of the workforce — are now considered “disposable” workers by their employers. This figure, drawn from an MIT survey of 6,000 employees, reveals the culmination of a forty-year economic transformation that is accelerating with artificial intelligence.
Paul Osterman’s MIT Sloan study demonstrates how the American economy has methodically destroyed the social contract of traditional employment. Between structural precariousness and “AI washing” of layoffs, this research documents the transformation of one-third of jobs into temporary positions, without protection or career prospects. A shift that is reshaping the future of American work.
The Essentials
- 35% of the American workforce (50+ million people) are classified as “disposable” according to the 2026 MIT survey
- The transformation has accelerated since 2020: a growing share of newly created jobs are precarious or contractual
- AI serves as justification for 42% of restructuring according to surveyed executives
- The most affected sectors include retail, food service, and logistics
Forty Years of Dismantling the Social Contract
The precarization of American employment did not begin with ChatGPT. It results from a systematic management strategy initiated in the 1980s, which the MIT study documents with precision. In the 1980s, the vast majority of stable jobs with social benefits characterized newly created positions. This proportion now constitutes a minority of new job creation.
Paul Osterman identifies several distinct phases of evolution. A first period corresponds to the rise of outsourcing and external consultants. Companies transform their permanent employees into service providers to reduce social costs. A later phase generalizes short-term contracts and the platform economy. Uber, TaskRabbit, and their equivalents normalize the absence of social protection.
The current phase, which began in 2020, uses artificial intelligence as justification for restructuring. The MIT survey reveals that 42% of executives invoke automation to justify job cuts, even when no technology is actually deployed. A phenomenon the study calls “AI washing” — the use of AI as a pretext for purely financial decisions.
The Anatomy of Disposable Jobs
The MIT study defines a “disposable” worker according to four objective criteria: absence of an indefinite-term contract, no employer-provided health insurance, hourly wages that vary from week to week, and the possibility of layoff without notice or severance.
On these bases, 51.2 million Americans fall into this category in 2026. The phenomenon concentrates three sectors. Retail employs millions of people on precarious contracts, representing a significant portion of its workforce. Food service counts a comparable number of disposable workers. Logistics accounts for millions as well.
But precariousness is spreading to skilled employment. Information technology now has 2.3 million consultants and freelancers versus 1.8 million permanent employees. Legal consulting employs 890,000 independent lawyers versus 1.2 million permanent ones. Even higher education relies heavily on adjuncts: 67% of university courses are taught by non-tenured instructors.
This precarization particularly affects young graduates. According to the study, 68% of those ages 25-35 hold insecure jobs, compared to 31% of those ages 45-55. A generational inversion that compromises access to home loans and wealth accumulation for an entire generation.
AI as Pretext for Restructuring
Artificial intelligence now fuels 42% of layoff justifications according to the MIT survey. But analysis of concrete cases reveals a striking gap between discourse and reality. Among numerous companies that invoked automation to eliminate positions in 2025, only a minority actually deployed AI tools in the months that followed.
This instrumentalization follows a now-perfected script. Executives announce a necessary “digital transformation” requiring a “skills reorganization.” Layoffs occur within three months. Technology deployment, when it happens, occurs 12 to 18 months later with reduced staffing and redefined missions.
The example of RetailCorp, a store chain anonymized in the study, illustrates this mechanism. The company lays off 2,400 employees in January 2025 to “automate the customer experience.” Six months later, no AI tools equip the stores. The eliminated tasks are simply redistributed to remaining employees, whose workload increases by 34%.
This strategy relies on public acceptance of technological disruptions. Invoking AI legitimizes restructuring that would be contested if presented as purely financial. American tariffs destroy more jobs than they create, but automation benefits from a symbolic capital that facilitates job elimination.
Refuge Sectors are Eroding
The MIT study documents the erosion of the last bastions of stable employment. The federal civil service, traditionally protected, now employs 23% contractors compared to 8% in 2010. Local governments are massively outsourcing: 41% of municipal services are contracted to private providers in 2026, compared to 22% in 2015.
Public education follows this trend. Substitute teachers represent 31% of staff in American schools, a proportion that has doubled since 2020. This precarization compromises educational continuity and fuels flight to private schools among families who can afford it.
Even healthcare, a sector in permanent tension, is precarizing its jobs. Public hospitals employ 43% temporary nurses, compared to 18% in 2019. This proportion climbs to 67% in rural states where facilities struggle to recruit. The human cost is documented: burnout rates reach 72% among precarious healthcare workers compared to 38% among permanent ones.
Only highly specialized jobs resist this trend. Scientific research, cutting-edge engineering, and finance maintain high levels of stable employment. But these sectors represent less than 8% of the total workforce and require qualifications that the majority of workers lack access to.
Macroeconomic Consequences are Emerging
This massive precarization generates systemic effects that the MIT study is beginning to quantify. Household consumption has stagnated since 2022 despite sustained economic growth. Disposable workers consume significantly less than their counterparts on permanent contracts with equivalent income, out of caution regarding uncertainty.
Access to credit is mechanically contracting. American banks are hardening their criteria: 73% of mortgage loans now require a permanent contract of more than two years. This restriction particularly affects young workers, of whom only 32% meet this criterion compared to 67% in 2015.
Professional training is collapsing. Employers invest 68% less in training their precarious employees. This progressive under-qualification threatens long-term competitiveness, particularly against a China that is industrializing the humanoid robot with a stable and trained workforce.
Public Initiatives Remain Limited
Facing this transformation, policy responses remain fragmented. The Biden administration launched a “Future of Work Council” in 2024 with $2.3 billion over three years. But this budget represents less than 0.05% of GDP, compared to 1.2% devoted to employment policies in Germany.
A few states are experimenting with innovative solutions. California is testing an intermediate status between employee and independent contractor, offering minimal social protection to platform workers. New York is developing training for emerging professions. Washington is financing retraining for employees laid off due to automation.
These initiatives remain marginal compared to the scale of the phenomenon. In Europe, SMEs await public guarantees to robotize, creating an alternative model for accompanying technological transitions. But the United States maintains its liberal approach, leaving the market to organize this transformation.
Toward a New Definition of Work
The MIT study concludes with a prospective question. Does this massive precarization herald the emergence of a new economic model or the disintegration of the current system? Paul Osterman would lean toward the first hypothesis, observing the progressive adaptation of workers to this permanent uncertainty.
Young generations are developing unprecedented strategies. Multi-activity, self-funded continuous training, and the creation of dense professional networks replace the security of traditional employment. This adaptation does not eliminate precariousness, but changes its nature.
It remains to be determined whether this transformation creates more opportunities than it destroys. The 50 million American disposable workers are experimenting, whether they like it or not, with a new relationship to work. Their capacity to adapt will determine the viability of this economic model for the decades to come.