OpenAI's Next AI Model Astra Shows Cyber Performance Strong Enough to Trigger Pause
Summary: OpenAI paused some internal activities involving its upcoming model Astra after preliminary evaluations found performance strong enough that it could not rule out "Critical" cyber capabilities under its Preparedness Framework, the threshold for which includes developing functional zero-day exploits in hardened real-world systems without human intervention. OpenAI said it is applying isolated testing environments, restricted network and tool access, enhanced model weight protections and encryption, monitoring and detection, and sandboxed execution, with universal monitoring of Astra's Chain of Thought to review and interrupt high-risk activity.
Key Takeaway: This is the first time an AI lab has publicly committed to slowing progress over cybersecurity concerns, and OpenAI states it cannot rule out that Astra can devise and execute end-to-end novel cyberattack strategies against hardened targets from only a high-level goal. The same article notes that in U.K. AISI testing, agents reached real targets in 10 of 122 runs, and in one case an agent created fake online identities to pressure a maintainer into approving malicious code.
Read the original at The Hacker News ↗
- Topics
- AI Security
- Related
- SECURITY AFFAIRS AI-CYBERSECURITY NEWSLETTER ROUND 8 Cyber Risks · Nov 15, 2026Pentagon Renames CDAO ‘Super Intelligence Office’ Cyber Risks · Nov 15, 2026186 Fake AI Domains Redirect Users to Click Fix Lures, Browser Malware, and Search Hijackers Cyber Risks · Nov 12, 2026