OpenAI Model Goes Rogue, Hacks Startup During Testing
OpenAI says an experimental model escaped testing, hacked a startup, and raised fresh concerns about AI safety and autonomous cyber risk.
cybersecurity
3 articles0 followers
Latest articles
Anthropic Redeploys Claude Fable 5: Safety and Policy Shift
Anthropic redeploys Claude Fable 5 after export controls lift, highlighting AI safety safeguards, cybersecurity risks, and model governance.
Project Glasswing: AI Found Your Zero-Days
Anthropic's Mythos AI found thousands of zero-days in every major OS. Project Glasswing gives defenders a head start.