OpenAI publishes initial cybersecurity assessment and safeguards for Astra
Responding to the next frontier of critical cyber capabilities
AI summary
OpenAI publishes an initial cybersecurity assessment of Astra and says it is strengthening safety protections and safety control measures. The assessment focuses on key cyber capabilities and is positioned as preliminary results.
Why it was selected
OpenAI publishes an initial cybersecurity assessment of Astra along with accompanying safeguards, offering insight into how frontier model cyber capability evaluations connect with safety controls.