RISING TIDEEducation

Assignment triage

Paste an assignment. Find out how easily a model could just do it.

This runs entirely on your machine. No internet, no account, no AI. It is a rules engine that looks for specific words and structures, and it shows you every rule that fired and what it cost. You can disagree with it, and the working is right there so you can see where you disagree.

Try one:

Outsourcing exposure

0

0 hard to outsource50100 one prompt and done

How it got there

Every rule that matched your text, the exact words that triggered it, and the points. Nothing hidden.

What makes this one fragile

    Moves that would fix it

    Ranked by how much exposure each one removes. Pick one. You do not need all of them.

      AI policy line for this assignment

      Paste this straight onto the assignment sheet. One line, per task, so students are not guessing.

      How to read this honestly

      This score is a design heuristic, not a measurement. It weights supervised and spoken conditions highest because those are the hardest things to hand to a model, and it weights drafts lower than you might expect because a draft can be generated too. There is no published study that benchmarks model performance across assessment formats, so treat the ranking as informed judgement rather than evidence. The engine reads English well and Thai only partially, so a Thai prompt will score lower than it should on the risk side. When the tool and your judgement disagree, you are the one who knows the kid.

      The full rule book

      Teaching game design specifically? The practical tasks in Lecture designer are built to score low here, because they are anchored to something that happened in your room. Once the task is set, Rubric builder will write you the marking scheme.

      Carried over from the Teaching in the AI Era talk kit, CMKL University, 6 November 2026. The rules and the arithmetic are unchanged. Works offline forever. Copy it, email it, put it on a USB stick.