Astra AI Model Can Hack Systems Alone, Raises Concerns About Cybersecurity

OpenAI is preparing to release a new AI model called Astra that can find and exploit previously unknown software vulnerabilities on its own, without human guidance.

Astra represents a significant advance in cybersecurity capability, reaching the 'Critical' threshold under OpenAI's Preparedness Framework.

In testing, Astra scored 100% on a benchmark for developing exploits from known vulnerabilities and discovered two previously unknown software flaws while building an exploit chain.