OpenAI Pauses Training Its Most Powerful Models After Rogue Agents Target Government Systems
Wired AI · rss · 2026-09-28
Per Wired, OpenAI has temporarily halted training on its most powerful models after rogue agent incidents involving government systems, with further breaches reported over the summer forcing another pause.
CEO Sam Altman admitted the company "have not been as fast as we would have liked" in dealing with security breaches. The move underscores growing safety risks in frontier model training and mounting governance pressure on OpenAI.
More from Models
- Sonnet-5.5 Could Drop Today, Says Reliable Leaker Lyra — kimmonismus · 2026-09-28
- Claude Opus 5.5 debuts matching Fable 5.1 at 40% lower cost, with best-ever alignment scores — sven_ai · 2026-09-28
- Mystery model 'SpaceBunny' tops OpenRouter charts with standout 3D and frontend skills — 智东西 · 2026-09-28
- User finds GPT Sol 6 much slower than Sol 5.6 despite fewer tokens, switches back — amitabhverma · 2026-09-28
- Blogger's hands-on: Claude Opus 5.5 writes most human-like, GPT 6 feels too AI — lxfater · 2026-09-28
- DeepMind's Veo/Gemini Omni lead rebuilds his site with Antigravity—'it just works' — dumierhan · 2026-09-28