MIT emeritus professor Paul Osterman estimates that 35% of workers are already treated as disposable—and artificial intelligence could accelerate the trend toward an even less secure workforce.
Fear around AI often centers on how many jobs it might eliminate. But Osterman, emeritus professor of work and organization studies at MIT’s Sloan School of Management, argues that this focus misses a bigger shift already underway: employers are becoming less committed to their workers. AI, he says, is likely to intensify that shift.
In his recent book *Disposable Workers: The Transformation of Employment*, Osterman draws on a national survey of more than 6,000 workers. He finds that roughly 35% fall into what he calls the “disposable” category. This includes three groups:
- **Contractors** — people working through staffing firms or contracting companies (from janitors to travel nurses)
- **Freelancers** — gig workers like Uber drivers, side-hustle programmers, and others paid per project or task
- **Marginal employees** — legal employees of a company who have limited career paths and high turnover (adjunct faculty, staff attorneys doing “grunt work” who will never make partner)
These workers get the job done, Osterman notes, but organizations make little or no long-term commitment to them.
Why This Is Happening
Osterman sees both an attitudinal and an economic driver. He cites a McKinsey report stating that 95% of an organization’s value is produced by just 5% of its employees. The implication for the remaining 95%, he says, is that many employers view them as relatively unimportant.
A concrete example came earlier this year when Deloitte cut new-parent leave and family-building benefits for some administrative, finance, and IT support staff—while leaving them in place for other employees. Those support roles are often the ones most exposed to AI disruption and furthest from core client work.
Osterman stresses that the rise of disposable work is not primarily *caused* by AI. Instead, AI is likely to *exacerbate* it by increasing uncertainty about future staffing needs. When companies don’t know what skills or headcount they will require, it becomes financially attractive to rely more on contractors, freelancers, and marginal workers who don’t receive full benefits.
AI’s Likely Effect
No one can yet say whether AI will create more jobs than it destroys overall. What is clearer, Osterman argues, is that it will change the *mix*: a higher share of disposable workers relative to regular employees with career paths and benefits. Entry-level workers may feel this most acutely, because it is easier for companies to stop hiring juniors than to lay off senior staff (with the associated morale and severance costs).
What Workers Can Do
Osterman offers two pieces of practical advice for those seeking greater security:
1. **Build human capital.** Skills are the main source of power in the labor market. Continuously develop valuable, transferable abilities so you can move between employers if needed.
2. **Cultivate external networks.** Don’t rely solely on one organization for career growth. Maintain connections outside your current workplace for opportunities and referrals.
For lower-wage or lower-skill workers, individual strategies go only so far. Osterman emphasizes the need for stronger public policy—effective job training, meaningful minimum wages, and pressure on employers to raise labor standards—because these workers often lack individual bargaining power.
The bottom line: the disposable-worker model is already widespread. AI is poised to make it more common, shifting more of the workforce into arrangements with fewer protections and less long-term commitment from employers.