From Creator to Auditor: The Quiet Deskilling of the Professions

Ask a junior lawyer, analyst, or engineer what they did today and you may hear a new kind of answer: they checked the machine's work. They did not draft the memo, build the model, or write the function from a blank page. They reviewed, approved, and moved on. Faster. Easier. And something is quietly going missing.

The headline fear about AI and work is replacement. The subtler story, and maybe the more important one, is transformation: the shift of skilled professionals from creators into auditors.

Competence lives in the messy middle

Expertise is not built by producing finished answers. It is built in the frustrating middle — organizing messy data, chasing an anomaly, holding three half-formed arguments at once until one resolves. That struggle is not waste. It is the process that lays down judgment. Outsource the middle, and you keep the output while losing the apprenticeship that used to come with it.

The risk, then, is a profession full of people who can evaluate a result but could no longer generate one — and who therefore, over time, get worse at evaluating, because evaluation rests on the same deep knowledge that doing the work used to build.

Designing against the drift

This is not inevitable, and it is not an argument for throwing the tools away. It is an argument about how the tools are built. An interface can require a practitioner to commit to an independent judgment before it reveals its own. It can offer to act as a devil's advocate or a proofreader rather than a wholesale author. It can track when a person's work has tilted entirely into auditing and flag the drift before the skill is gone.

The goal is a working relationship in which the machine sharpens the professional instead of slowly standing in for them. Keeping that distinction is one of the most concrete fights inside cognitive sovereignty — because it is happening, right now, in nearly every skilled job at once.