Ethan Mollick: AI detector evasion tools produce bad writing

eldonredwards · x · 2026-08-19

Ethan Mollick retweeted a post about Deft, an AI writing lab designed to generate text that evades AI detectors. He critiqued that the fundamental issue with such tools is that they produce bad writing.

The original post claimed Deft achieves 86% "fully human" ratings on Pangram and advised users to add more detail to prompts or change styles to improve results—strategies that often result in stilted or low-quality prose.

Related event: New AI Lab Deft Launches Beta Writing Model Claiming 86% Human-Rated Output(7 posts)→

Original post →

More from Apps

Apps channel →