Wire · opportunities
OpenAI rates Astra 'critical,' curbs hacking features before release
◆ Sectors
◆ Geography
◆ Source
◆ Verified
Fusion42 · 2 September 2026 · Fusion42 review
OpenAI's AI model Astra achieved a 'critical' cybersecurity rating for its ability to autonomously discover and exploit zero-day vulnerabilities, prompting OpenAI to restrict its hacking features and implement stronger safety measures prior to public release.
This Wire brief sits within Fusion42's coverage of AI Frontier Models, and 4 sources have reported it between 1 Sep 2026 and 2 Sep 2026.
◆ ◆ The Wire takeaway
You must treat AI cybersecurity tools as potential threats alongside technologies to secure. OpenAI’s move forces founders building AI cyber tools to prepare for stricter access controls and compliance hurdles starting now.
◆ Coverage
4 sources · first reported 1 Sep 2026 · latest 2 Sep 2026
◆ Related on Wire
◆ Topics