OpenAI tightens security for future Astra model

OpenAI has introduced stringent security protocols for its unreleased AI model Astra after internal tests show it could reach a 'Critical' risk level in cybersecurity, capable of autonomous zero-day exploit development and attack methods. Development activities that do not meet new security standards are paused, with testing to continue under tighter controls and cooperation with government and AI security groups.

This Wire brief sits within Fusion42's coverage of AI & ML, and 3 sources have reported it between 7 Aug 2026 and 10 Aug 2026.