Researchers question why Anthropic doesn't report misalignment incidents like OpenAI

sjgadler · x · 2026-11-19

Safety researcher sidbaines asks why Anthropic doesn't publish misalignment incident reports the way OpenAI does, arguing such disclosure would greatly benefit loss-of-control risk research, isn't very costly, and that Anthropic almost certainly has incidents to report. The post directly tags Dario Amodei, urging Anthropic to "race to the top" on transparency.

Original post →

More from Safety

Safety channel →