Notes on AI, integrity,
and the classroom.
Written for educators and school IT by the team building Themisto EDU. What we see in real schools, what works, and what quietly doesn't.
How to Tell If a Student Used AI, Without Trusting a Detector
AI detectors flag honest students and miss the ones who edit, so they cannot be the answer. What holds up is a change in process: knowing how the work was made, not guessing from the finished text. Here is what actually tells you, and what to stop relying on.
An AI Syllabus Statement That Actually Holds Up
A one-line ban on AI is impossible to enforce and unfair to apply. What works is a syllabus statement that says exactly where AI helps, where it does not belong, and what happens either way. Here is the structure, the language, and a version you can adapt.
AI Detectors Keep Accusing Honest Students. The Numbers Say This Will Keep Happening.
Detection scores look like evidence and behave like guesses. What the research actually shows about false positives, who gets flagged most, and what schools should require before any accusation.
Parents Are Asking About AI in the Classroom. Here Is a Straight Answer.
Parents want to know two things: is my child using AI to cheat, and is my child's data safe. Vague reassurance does not land. Here is the plain-language answer schools can give, and the approach that makes the answer true.
Blue Books Are Back. Here's What the Handwriting Comeback Fixes, and What It Quietly Doesn't.
Universities are reporting big jumps in blue book exam sales as professors retreat to handwritten, in-class assessment. It's a rational move. It's also not a plan.
How Students Actually Use ChatGPT. It's Not All Cheating, and That's the Hard Part.
Nearly nine in ten students now use generative AI for assessments. The same tool tutors one kid and does the homework for the next. Policy that can't tell the difference punishes the wrong behavior.
We Watched a School Ban ChatGPT. Cheating Went Up.
Blocking AI sites feels decisive and looks good in a board meeting. Then students switch to their phones, the tools rebrand weekly, and teachers lose the only visibility they had. Here is what actually happened, and what works instead.
Tutoring or Cheating? Same Tool, Same Kid, Different Question
The whole academic integrity debate collapses into one observation: the tool is neutral, the intent is not. 'Explain photosynthesis' and 'write my essay' go to the same chatbot. Telling them apart is the actual job.
Write Your AI Policy Like a Teacher, Not a Lawyer
Most school AI policies are three pages of legalese nobody reads, enforced by nobody, helping no one. The good ones fit on an index card. Here is how to write one, with examples you can steal.
The Integrity Conversation Is Easier When You Have the Receipt
Accusing a student of AI cheating on a hunch destroys trust whether you're right or wrong. The conversation changes completely when you can open a dashboard and talk about a specific prompt instead of a suspicion.
1:1 Devices Made AI Everyone's Problem. A Playbook for School IT
When every student has a laptop, every student has a tutor, a ghostwriter, and a homework machine. A practical rollout guide for IT teams who just inherited the AI problem: what to deploy, what to skip, and what breaks.