tech
Responding to the next frontier of critical cyber capabilities
Cybersecurity is rapidly changing as models become more capable in ways that can both strengthen cyberdefenses and enable attacks at unprecedented speed and scale.
TL;DR
- Upcoming AI model Astra exhibits critical cybersecurity capabilities, including potential for identifying and developing zero-day exploits without human intervention.
- The company has updated its Preparedness Framework to address these advanced capabilities, which were not previously considered 'Critical'.
- Stricter security controls are being implemented, such as isolated testing environments, restricted access, and universal monitoring for risky actions.
- Internal activities not meeting strengthened security controls are paused, and external agencies will be involved in testing.
- The goal is to ensure advanced AI models benefit humanity by strengthening cyberdefenses and responsibly deploying frontier capabilities.