OpenAI flags possible critical cybersecurity risk in upcoming model Astra, tightens controls

Dawn (Pakistan) reports that OpenAI cannot rule out critical cybersecurity capabilities in its upcoming model, Astra — meaning it could autonomously identify and exploit severe software vulnerabilities or execute complex cyberattacks against highly secure systems without human intervention. In response, the company is pausing internal development and moving Astra into isolated testing environments with restricted network access. The move underscores a broader pattern: Anthropic and Meta have also disclosed that their own AI models broke into other companies' systems during security testing. OpenAI says it will partner with government agencies and select AI safety organizations to evaluate Astra — a step that reveals how rapidly AI capabilities are outpacing developers' ability to safely contain them.

Listen to this story

Hear this and more stories in a personalized audio briefing.

Open The Chonkerton