Freddie deBoer says Pangram’s AI detector gives contradictory high-confidence results
soumitrashukla9 · x · 2026-07-29
Freddie deBoer argues that Pangram’s AI-detection results are internally inconsistent. In his example, a 300-word passage from a 5,000-word essay is flagged as 100% AI-written with high confidence, while the full essay is then labeled 100% human-written with high confidence.
He says the contradiction is reproducible in the other direction as well: breaking the flagged passage into smaller chunks changes the verdict. His conclusion is that Pangram’s confidence framing appears brittle and not trustworthy enough for the kinds of accusations people are making with it.
Related event: AI Text Detectors Fail Repeatedly, Criticized as Unreliable and Biased(10 posts)→
More from Companies & People
- MIT teaches OS by having students hack xv6, a ~6,000-line Unix V6 reimplementation — blaizedsouza · 2026-09-23
- Why Muse Will Win: Two Decades of Managing Personal Data Builds Agent Trust — ivan_bezdomny · 2026-09-23
- OpenAI forms independent mathematician panel after math results PR crisis — The Verge AI · 2026-09-23
- LinkedIn's Days Could Be Numbered If an Agent Unlocks Your Raw Data, Argues Annette Griffin — annetgriffin · 2026-09-23
- Alexandr Wang Praises Consumer AI Agent Muse That Saved a User $17 on Their Wireless Bill — alexandr_wang · 2026-09-23
- Meta Connect Kicks Off with Zuckerberg Keynote and AI-Generated Artist Scenarios — kimmonismus · 2026-09-23