Every Output Is a First Draft: Teaching Students to Interrogate What AI Hands Them

Means treats evaluation as a competency that has to be taught rather than assumed, and gives it a mechanism a class can run next week: her push-back protocol has students demand the evidence behind every claim in a model's answer, surface its assumptions, test it against a real application, and only then write their own analysis, submitting a record of what they checked, rejected and revised. She separates judgement, which is choosing among the options in front of you, from discernment, which is asking whether those were the right options at all, and she is explicit that the thing being graded is the interrogation and not the output. It is the shortest route from the sentence teach them to think critically to something you can actually put on an assignment sheet.

Task

Teaching AI literacy

Time

12 min

Level

beginner

Cost

Free — Nothing to pay. A free account at most.

Tools

None in particular

Published

2026-08-10

Voice

Tawnya Means — Writes for whoever has to fund and staff AI change across a whole college, including which pilots to run and which faculty to support first.