AI as the Latest Chapter in Capital's Historic Project to Discipline and Replace Labor
Source: Garrison Lovely. "The Bosses’ Oldest Dream." October 1, 2026. jacobin.com
The Gist
The author argues that AI isn't just a new technology—it's the latest and most powerful version of something bosses have wanted for centuries: workers who never complain, never organize, and never need to be paid. The piece traces this desire back to the 1800s and uses new economic data showing young workers in tech-exposed jobs are already losing employment to suggest this isn't science fiction anymore—it's happening now, and it's driven by the same profit motives that have always shaped how new technology gets used against workers.
Conclusion
The AI industry's drive to automate jobs is not a neutral technological development but the culmination of capitalists' centuries-long project—beginning with Babbage's calculating engines—to discipline, control, and ultimately eliminate dependence on human labor, and this process is already measurably harming young white-collar workers.
Premises
- The earliest computing devices (Babbage's calculating engines) were explicitly designed as tools of labor discipline, building on surveillance and control techniques first developed on slave plantations and later applied to industrial workers.
- Automation has historically served bosses not just through cost savings but by weakening workers' bargaining power and undermining labor organizing, as seen in Amazon's use of algorithmic management to surveil and suppress protected organizing activity.
- Contemporary AI industry leaders and corporate surveys reveal explicit intent to automate jobs as quickly as possible, with over 40 percent of large employers in a 2025 WEF survey planning workforce reductions due to automation by 2030.
- AI companies are generating unprecedented wealth with dramatically fewer employees than prior generations of top firms (e.g., OpenAI, Anthropic, xAI showing $140-167 million in value per employee versus $3.5-30 million historically), suggesting capital is substituting for labor at an accelerating rate.
- A rigorous August 2025 Stanford economics study using granular payroll data found a 16 percent decline in employment for 22-25 year-olds in the most AI-exposed fields within just over a year, providing early empirical evidence of AI-driven job displacement that broader government datasets (like CPS) were too coarse to detect.
- Economic modeling (Korinek and Suh) shows that even if AI automation drives substantial GDP growth, wages would collapse as the economy approaches full automation, undermining the traditional assumption that technological disruption always creates offsetting new jobs.
- This AI-driven dynamic is intensifying already historic levels of wealth inequality, last seen in the 1920s, risking an economic order where one's position depends entirely on wealth accumulated before their labor becomes obsolete.
Assumptions
- The historical motives behind early computing (labor discipline) meaningfully explain or characterize the motives and effects of today's AI industry, despite major technological and contextual differences.
- Expressed intentions by CEOs and employers to automate jobs will translate into actual technical capability and successful implementation.
- The Stanford payroll study's findings on young workers in AI-exposed fields are primarily attributable to AI adoption rather than other confounding factors (e.g., macroeconomic conditions, interest rates).
- Current trends in AI capability and corporate behavior will continue or accelerate rather than plateau or reverse.
- Wealth concentration metrics (value per employee) are a valid proxy for labor displacement rather than reflecting other factors like capital intensity or market dynamics unique to tech/AI firms.