The Signal · Column 021

AI in hiring does not remove judgment, it relocates it

Andrei van Noordt · 8 August 2026 · 1 min read

When the United States Postal Service, one of the largest employers in the country, says it is exploring AI in its hiring while reassuring workers that humans will not soon be replaced, the reassurance is the part worth reading. No organization feels the need to promise it will not replace people unless replacement is what everyone quietly assumed the tool was for.

The reading most people miss is that AI in hiring does not remove human judgment from the process. It moves it. When a machine handles the top of the funnel, the screening, the ranking, the first pass, the judgment that used to be spread across a room of recruiters gets concentrated into the few people who can tell when the machine is wrong. The screening work gets cheaper. The overriding work gets rarer, and rarer has always meant more expensive.

The promise that no one will be replaced is really an admission that the hardest part of hiring was never the screening.

This is the repricing to watch. Every hiring function that puts AI at the front of the funnel still needs someone whose feel for a candidate is good enough to catch what the model filters out. The unusual background, the career gap that hides a strength, the strong hire that a pattern match would have discarded. Those people were valuable before the tools arrived. Now their judgment governs a system that touches every single applicant, which makes them the most leveraged hire in the department, not the most expendable.

For hiring leaders, the practical read is simple. If you deploy AI to screen, invest in the same breath in the senior recruiter or hiring manager who can audit it and overrule it, because a fast funnel pointed at the wrong candidates does not fix your hiring. It only makes bad hiring more efficient.