ExperimentOctober 10, 2026
Does the Human Check flag my own published posts?
A rules-based scanner will flag the AI-assisted posts already published on this site.
- Kind
- Eval
- State
- Done
- Run on
- 2026-09-27
01
Why I ran it
To find out whether the score means anything before relying on it.
02
Setup & method
Ran the deterministic Human Check scan over every published entry with enough prose.
03
What happened
68 posts scored an average of 94, and none were flagged, even though much of the content was written with AI assistance.
04
Result (one line)
0 of 68 posts flagged. The scanner is lenient on long technical prose.
05
What surprised me
It works as a style linter for obvious tells, not as a judge of how something was written.
06
Limits — what this does not prove
The weights are hand-tuned, not calibrated against labelled examples. A high score means few known tells, nothing more.