Conversation

Notices

1. Embed this notice
   Mignon Fogarty (grammargirl@zirk.us)'s status on Tuesday, 13-May-2025 23:18:41 JST Mignon Fogarty

   This whole article about who can and can't detect AI writing is fascinating!

   The average person essentially does no better than chance, but a panel of 5 people who use AI for writing every day got only 1 in 300 wrong.

   The screenshot describes how they look at the writing differently.

   https://arxiv.org/pdf/2501.15654

   • Haelwenn /элвэн/ :triskell: likes this.
   • Embed this notice
     Mignon Fogarty (grammargirl@zirk.us)'s status on Tuesday, 13-May-2025 23:28:12 JST Mignon Fogarty

     in reply to

     Here's an absolutely wild detail!

     To use paired articles on the same topics, they started with human-written articles and then had AI generate articles of the same topic and length, so it had to make up names for people in the stories, and it used the same names ... a lot:

     "Annotator 1 is the only one to pick up on 'AI names': in fact, 63.3% of GPT-4O and 70% of CLAUDE-3.5-SONNET articles include either the name Emily or Sarah."

     Haelwenn /элвэн/ :triskell: likes this.