Pangram AI Detector Gives Conflicting Results, Sparking Debate

An X user questioned whether lukedrago's post was AI-written, showing Pangram 4.1 flagging a 56-word excerpt as 66% likely AI. The author responded by running the full 258-word text through the same detector, which returned 100% human-written, sparking debate over AI detector reliability.

2026-11-11 ~ 2026-11-11 · 2 related posts

• Author falsely accused of AI writing shows Pangram 100% human verdict — luke_drago_ · 2026-11-11
• Pangram flags human-written text as 66% AI, sparking detector false-positive debate — austinc3301 · 2026-11-11