OpenAI Documents Safety Failures in Long-Running AI Systems
OpenAI documents safety failures in long-running AI agents, naming five risk classes and new safeguards as autonomous systems act for hours.
-
-
Why Robots Still Face a Dangerous Safety Crisis
Why Sonair CEO Knut Sandven argues that smarter robots still cannot guarantee factory safety.
-
AI Pretraining Poisoning Exposes a Shocking New Gap Safety Can’t Close
Pretraining poisoning survived fine-tuning and RLHF in a July 16 study, exposing a safety gap that post-training checks may not close.
