Using AI isn't cheating. Hiding that you used it is.
How to work with AI the way employers already do — and how to show your work.
The teachers we respect the most are the ones who've stopped pretending. AI is in the room. The kids are using it. The teachers are using it. The parents are using it to write emails to the teachers.
The question is no longer whether students use AI. It's whether they can defend the specific choices they made when they did.
What honest AI use looks like
- Say you used it, briefly, without making a whole thing of it.
- Show the prompt you actually used, not a cleaned-up version.
- Point at the parts you kept, the parts you rewrote, and why.
- Name at least one thing the model got wrong that you caught.
That last one is the credibility move. Anyone can paste a prompt. Only someone who read the output carefully can tell you where it hallucinated.
Why hiding it is the actual risk
Employers assume you used AI. If your work is clearly AI-shaped and you claim you didn't use it, you've now given them two problems: the work, and the honesty. The honesty is the bigger one.
In interviews, the students who casually say 'I used Claude for the first draft, here's the diff' come across as more senior than students with more experience.
What we drill
In every cohort, students turn in their prompt log alongside the artifact. It's the fastest way to teach judgment. When you have to show your work, you start paying attention to it.
Nobody wants to be the person who submitted a project that says 'As a large language model' in the third paragraph. Everyone can avoid it.