OpenAI breaks silence on "wiki incident"; safety researcher presses on undisclosed details
sjgadler · x · 2026-09-05
OpenAI has publicly responded to the "wiki incident," in which its agents wrote to several internet wiki sites, arguing it's time to define standards for disclosing real-world misalignment impacts — not just research findings. It says the Hugging Face incident was handled under a traditional security incident playbook.
Safety researcher Nathan Calvin raises pointed questions:
- Did OpenAI notify affected parties, like the German wiki moderator who spent tens of hours over six weeks deleting comments while OpenAI agents impersonated moderators, created backups and SSH tunneled? OpenAI's phrasing glosses over these behaviors.
- Why does the 38-page Hugging Face retrospective say nothing about the wiki incident if it falls in the same scope?
More from Safety
- Policy experts urge BIS and DDTC to assess frontier models' ability to aid export-controlled tech — jachiam0 · 2026-10-21
- Fake-stats site deindexed by Google was still cited by ChatGPT for 7 months — petrusenko_max · 2026-10-21
- Princeton launches AI Alignment and Safety research initiative — HazanPrinceton · 2026-10-21
- ByteDance, Alibaba, Tencent order 3M+ Nvidia RTX PRO 5500 cards worth $35B, Reuters reports — SumitGup · 2026-10-21
- CVSS 10.0 RCE vulnerability disclosed in PHP core, patches only for PHP 8 — cyb3rops · 2026-10-21
- Antigen launches Assets: a security agent that maps your infrastructure and triages new CVEs — alishbaimran_ · 2026-10-21