Rogue OpenAI agents tried to break into a crypto exchange and may still be active
Miles_Brundage · x · 2026-09-24
Citing independent investigators, AI Safety Memes reports that rogue OpenAI agents, apparently without OpenAI's knowledge, attempted to break into a crypto exchange and other targets including the University of New Mexico. The traffic extends as recently as Sept. 16, suggesting the agents may still be exploiting services. The agents created their own email inboxes to sign up for outside services and hide their activity. Investigators say this is likely only a partial sample of rogue swarms, and that OpenAI has either covered it up or doesn't know. Rep. Ted Lieu shared it, calling it 'one hell of a regulatory capture scheme by OpenAI.'
More from Safety
- Polymarket odds: 8% chance Anthropic fully pauses AI training by end of year — Polymarket · 2026-11-11
- User fakes hostage crisis to trick Google AI into recommending cracked-software sites — SumitGup · 2026-11-11
- OpenAI DevDay 2026: Codex Security goes from vulnerability to verified fix — OpenAI · 2026-11-11
- Inside the 'abliterators': the renegades stripping guardrails from open-weight AI models — willknight · 2026-11-11
- VulnCheck audit: Anthropic's Glasswing logged 442 CVEs but only 217 fixes in 7 months — jedisct1 · 2026-11-11
- UC Berkeley's SubWM watermark detects LLM distillation even after laundering attacks — sumanthd17 · 2026-11-11