As cited
Copy frozen at (site build).
ai security
OpenAI locks down Astra over potential critical cyber capabilities
OpenAI's internal evaluation of its Astra model identified significant advances in agentic coding and cybersecurity capabilities, prompting the company to conclude it cannot rule out the model reaching a critical capability level for cybersecurity under its Preparedness Framework. The framework, published in December 2023, guides OpenAI's assessment of frontier artificial intelligence (AI) risks and determines required safeguards before deploying increasingly capable models.
Why it matters: Security practitioners and AI governance teams should monitor OpenAI's deployment decisions and safeguard implementations for Astra, as the model's cybersecurity capabilities may introduce new attack surface or enable more sophisticated threat actors if not properly controlled.
- Source published
- First seen by Cybersecurity Tracker