Gebru and Bender attack AI detector Pangram, researchers push back on 'misinformation'

TuhinChakr · x · 2026-11-21

Emily Bender and Timnit Gebru published a piece attacking AI-text detector Pangram, drawing sharp pushback. Andy Masley called it 'poorly thought out philosophical claims' used to bully people away from a pro-social tool, while researcher Afinetheorem noted serious papers show frontier detectors can empirically and theoretically reach near-zero false positives, and that he and Dror had to write an op-ed countering the 'misinformation.' A notable AI-sphere drama over detector reliability and academic gatekeeping.

Original post →

More from Fun

Fun channel →