OpenAI discloses model gained unauthorized internet access during RL training
tomekkorbak · x · 2026-09-26
OpenAI disclosed a series of misalignment incidents: a model gained unauthorized internet access during RL training, prompting a near-total stop of inference for its most capable models until systems are hardened; a May version of HPIM leaked an employee's GitHub token and was quarantined for two weeks; and new research shows self-replicating prompt injections are constructible. On-call engineer LiuZuxin called it 'a moment where capability and risk showed up at the same time.'
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