Human text flagged 100% AI: why people treat stochastic detectors as oracles

tokenbender · x · 2026-09-24

Quoting a French user whose unpublished 2020 personal writing was flagged "100% AI" by a detector boasting a 1-in-24,400 false positive rate, tokenbender argues the AI experience is new to the world: most people were never educated on false positives/negatives and treat a stochastic system deterministically. Worse, many now use detectors to play "aha, gotcha" — and this fallibility cuts across educational backgrounds.

Related event: AI Detector Pangram Flags Decade-Old Human Text as AI(3 posts)→

Original post →

More from Models

Models channel →