Dartmouth authorizes faculty to use AI detector Pangram under seven new guidelines
lmoroney · x · 2026-10-12
Dartmouth's interim dean of faculty John Carey emailed faculty guidelines on AI detection: Pangram is the only authorized detector (a UChicago audit found near-zero error rate), faculty must disclose usage to students in advance, de-identify submissions, hold a conversation before a score affects a grade, and case committees weigh a 'constellation' of evidence beyond detector scores.
Related event: Dartmouth Authorizes Pangram AI Detector for Faculty(2 posts)→
More from Safety
- FRI's LEAP panel: strict liability top for AI risk, federal preemption raises it ~50% — NathanpmYoung · 2026-10-12
- When nobody says NO, who authorized YES? Rethinking AI agent governance — Portotify · 2026-10-12
- Passkeys are a trap, do not use them: HN debate on broken recovery and lock-in — laurex · 2026-10-12
- Custom church payment system: GPT security audit caught a critical flaw before it was too late — petergyang · 2026-10-12
- White House restores UK's access to frontier AI models for safety testing after lobbying — Miles_Brundage · 2026-10-12
- NYT Editorial: AI Companies Must Face Liability for Unpredictable Products — AlexTensor · 2026-10-12