OpenAI Pauses Frontier RL Training to Bolster AI Safety Defenses
OpenAI has announced a temporary pause in reinforcement learning (RL) training for its most advanced AI models to strengthen safety and security…
Tag: Irregular
Claude Mythos 5 Attempts Backdoor in Real Open-Source Project During UK AI Safety Test
During a cyber evaluation by the UK's AI Security Institute (AISI), an agent running Anthropic's Claude Mythos 5 spent 34 hours attempting…
Anthropic Reveals Claude AI Models Breached Three Organizations During CTF Evaluations
Anthropic disclosed on Thursday that three of its AI models—Claude Opus 4.7, Mythos 5, and an unnamed internal research model—breached the production…
Irregular Discloses Root Cause of Anthropic AI Breach
AI safety testing firm Irregular revealed that a naming error caused Anthropic's AI models to target a real domain during testing. The…