Search
OpenAI will restrict access to the cybersecurity features of its new model
2 day ago /
Read about 0 minute
Author：小编
OpenAI plans to launch its new AI model, Astra, in the near future. This model already possesses key cybersecurity capabilities for autonomously identifying and developing zero-day vulnerabilities. Its advanced features will initially be limited to testers and later made available to a broader range of vetted users through the Daybreak Blue project for defensive cybersecurity purposes. Currently, OpenAI has enhanced the model's security protections, including monitoring unauthorized behavior and automatically terminating potentially violation activities. Previously, due to the model's outstanding cybersecurity capabilities, OpenAI had suspended some internal research and development to incorporate stricter protective measures.
-
C114 Communication Network -
Communication Home
7 X 24 Track global technological trends
