Digital labor
What is Digital labor?
Software that does real work a person used to do, not a single task but chunks of a whole process. IBM frames today's version as AI agents that run multi-step jobs and hand back finished results for a human to review. The pitch is capacity without added headcount. The open question is oversight and accountability.
Why it matters
Digital labor reframes AI from a tool a person picks up into a worker a person manages, and that change in language matters more than it first sounds. Once software handles parts of a role rather than assisting with a task, the questions turn organisational. Where does it sit on the org chart, who is accountable for it, how does its work get counted. It also forces issues the tool framing lets a company dodge: what happens to the people whose tasks it absorbs, how you trace a decision the software made, who carries the blame when it goes wrong. Vendors sell the upside hard. The harder half is building the responsibility around it that a human worker already brings with them.
In practice
Instead of buying a tool to help an analyst format reports, a company sets up a process where the software gathers the data, drafts the report and passes it to a person only for approval. That is digital labor: a slice of the job done end to end. The analyst’s role moves from producing to supervising, and the company has to define what “supervising” means before the volume grows.
