AI and work presents a fork: redistribution within the workplace, or redefinition of the workplace. A socialist analysis should distinguish between the two. Vivek Chibber’s recent piece in Jacobin doesn’t.1
Chibber tells a familiar story. Technology displaces workers in one sector, the economy grows, new jobs absorb them elsewhere. Aggregate employment stabilises. He opens with the spinner—handloom weavers displaced by mills—as proof. But spinners didn’t disappear. They became operators.
The replacement job wasn’t simply ‘a new job in a growing economy.’ It was the same worker subordinated to the machine: less autonomy, less craft knowledge, pace set externally rather than by judgment. Employment statistics held. The experience of work degraded. That distinction matters.
The deeper history isn’t just mechanisation but the collapse of the putting-out system. For generations, textile merchants dispersed production to cottage weavers. Workers controlled their pace, worked for multiple merchants, retained leverage because coordination was costly and production was distributed.
The factory didn’t win on productivity alone. It solved a control problem. Concentrate labour. Own the inputs. Let the machine set the tempo. The result wasn’t merely sectoral reallocation. It was a reorganisation of authority inside production. That’s the fork Chibber’s framework blurs.
Redistribution within the workplace is manageable. Tasks shift. Some jobs deskill. Organise labour. Expand welfare. Bargain over the new equilibrium. Redefinition of the workplace is structural. It alters who controls pace, discretion, and coordination. It changes not just what workers do but how work is constituted.
The more ambitious AI rhetoric points toward the second. Not mass unemployment, but dissolution of the employment relationship itself—dispersing cognitive tasks while using the algorithm to solve the control problem the factory once solved. You can hear executives reaching for it. What you can’t yet see is the substrate to make it work.
Agentic systems remain brittle outside narrow, well-specified domains. Workplace tasks are embedded in organisational context and tacit knowledge. Handoffs are frequent and unpredictable. Control at scale requires reliability across those seams. We don’t have that.
The present moment may therefore be less like the factory transition and more like capital using AI rhetoric to pursue restructuring goals it already had, with technology supplying marginal capability and strategic justification.
But the fork remains. Redistribution or redefinition. That’s the question.
- Chibber, Vivek, and Melissa Nascheck. “Is AI Coming for Our Jobs?” Jacobin, February 28, 2026. https://jacobin.com/2026/02/ai-technology-productivity-growth-job-loss. ↩︎