AI Raises the Floor, Rather than Simply Taking the Job

Whether AI will take jobs has become a familiar question. During the same period, other research findings have accumulated. Reports from call centres and customer support describe generative-AI assistance improving the performance of less experienced staff most. Veteran productivity changed little, while newcomers' service quality moved closer to theirs.

This is a story of raising the floor, not replacement. Rather than taking a person's place, AI quietly narrows gaps in experience, intuition and tacit knowledge. These findings do not automatically extend to every occupation. Effects vary greatly by industry and task; it is more honest to regard the phenomenon as one we have only begun observing. Still, this is a different context from the AI-threat narrative.

Expertise has traditionally been measured by the amount one knows: years worked, experiences accumulated and patterns internalized. If AI can perform part of that tacit work, the definition begins to shift. Knowing things becomes relatively less valuable than deciding what to ask, select or discard in a particular situation. Expertise moves from quantity toward the quality of judgment.

This may be welcome for individuals and small businesses. AI can bridge some of the accumulated know-how separating them from large organizations. A solo business may produce writing, analysis and proposals approaching an experienced professional's level from an early stage. The question becomes not possession of the tool but what to ask it, how to assess its output and where to put one's own name. The quality of a question starts distinguishing work more than sheer hustle.

Raising the floor has another side. If everyone's minimum rises, work once protected by years and experience quietly loses some value. Equalization also lowers entry barriers. When a passing grade becomes easier to reach, passing itself is worth less. Distinctive perspectives, responsibility and habits of judgment become important sooner. Improvement is both help and a quiet form of selection.

Asking whether AI takes jobs is clear but coarse. What may actually be happening is a gradual rewriting of expertise from accumulated experience toward judgment. Which parts of our own work are protected by quantity, and which by judgment? It is worth pausing to separate them.