Definition
Plain language
How often a detector is tricked into calling machine-written text human.
As stated in the literature
The fraction of AI-generated texts a detector misclassifies as human; higher means a more successful evasion attack.
Why it matters: It gives a single number for how easily a detector can be evaded, which is what people rely on to judge whether a detector is trustworthy.
For example, if a detector wrongly labels 30 out of 100 machine-written essays as human-written, its fool rate is 30 percent.
Heard on the show
“The number is the fool rate — the fraction of AI texts the detector wrongly labels human.”Episode 220 — Write Like It's 1923: The One-Prompt Trick That Beats AI Detectors