Your employees didn't become parrots because of AI. Your evaluation system did.

The AI Rollout Problem

Companies everywhere are anxiously pushing AI: hand out the accounts, tell everyone to work faster.

Here's what you get instead. Reports written by AI. Decks generated by AI. Employees who've become readers of AI output. And managers who can no longer tell what the person did and what the tool did.

Most people blame the AI. But that's not where the problem lives.

The Proxy That Used to Work

For years, organizations treated "a polished report" as a proxy for "this person thought hard and understood it." That proxy worked — because producing the report used to require the thinking.

AI cut the cord between the deliverable and the thinking.

So the proxy failed, in public. "Employees became parrots" is just the quiet truth finally surfacing: we were grading packaging, not judgment.

What to Ask Instead

So stop asking whether people are using AI.

Ask whether the quality of their judgment got better.

A Simple Test for Managers

One thing managers can do tomorrow: in a review, stop staring at the polished deliverable. Probe the thinking behind it. "Why did you land on this conclusion?" "What did you change from the AI draft, and why?"

The people who actually understood, and the people who are just reciting, separate themselves in about thirty seconds.