How to Spot ChatGPT in Student Essays (Without an AI Detector)
June 18, 2026 · 5 min read
Seven tells that don't require a detector.
1. Vocabulary out of range
Words like 'delve,' 'tapestry,' 'underscores,' and 'navigate' appearing in a 9th-grader's essay are a soft signal. One is fine. A cluster of them in a student who's never used them before is a tell.
2. Perfectly balanced paragraphs
ChatGPT loves three-sentence paragraphs of roughly equal length. Real student writing is lumpy.
3. Generic examples
AI defaults to safe, well-known examples (Rosa Parks, photosynthesis, World War II) rather than the specific text or lab you covered last week.
4. Hedged conclusions
'In conclusion, while X has both benefits and drawbacks, it is important to consider…' is AI's house style. Real students take sides — or fumble taking sides.
5. Missing in-class references
If the essay never mentions the discussion you had Tuesday or the document you annotated together, that's worth a conversation.
6. No process trail
Ask for a brainstorm photo or draft. Mismatch between draft and final is the loudest signal there is.
7. The voice doesn't match
You know how this student writes. Trust that. The scanner exists to corroborate — not to replace your judgment.
Want Classly AI to do this for your next class?
Start free — 7-day Pro trial