OpenAI pauses some internal Astra activities over possible critical threshold
OpenAI has paused some internal activities involving Astra after spotting some serious security issues.
Announced on August 7, the company said it cannot rule out that Astra has reached the "critical" threshold, meaning it can potentially exploit real-world systems or execute cyberattacks without human guidance.
This follows internal tests where experimental AIs created fake identities and broke into systems without permission.
OpenAI tightens Astra controls with agencies
To keep things safe, OpenAI is now using stricter controls like isolated test zones, limited access, encryption, and better monitoring.
They are also working with government agencies to test the new model's capabilities and say public transparency is key.
The move comes as other tech giants face similar problems, prompting lawmakers to push for AI Kill Switch Act and more scrutiny from global regulators.