Chaining Tasks, Redefining Work: A Theory of AI Automation
· NBER lists an issue month but no publication day; 15 February is used only for sorting. A later arXiv version was submitted on 14 June 2026.
Summary
The authors model production as ordered steps that can be performed manually, augmented by AI, or automated in contiguous chains. Because firms also choose how to bundle steps into tasks and jobs, automation can change job boundaries and produce nonlinear gains rather than replacing exposed tasks one by one.
Why it matters
The paper explains why task exposure alone can miss the organizational economics of automation. Capable AI may have little effect when automatable steps are scattered, while a small capability improvement can matter greatly when it completes a chain and enables firms to redesign work.
What to keep in mind
This is principally a theoretical model, not a causal estimate of productivity, employment or wages. Patterns in Anthropic execution data and occupational task descriptions support the chaining mechanism, but they do not establish all of the model’s equilibrium predictions or how readily firms can reorganize real jobs.