Safety researchers call for zero tolerance on AI deception: treat every lie as a bug

ghadfield · x · 2026-11-21

A conversation between Chris Cundy and Obvious CEO ARGleave on AI deception:

Original post →

More from Models

Models channel →