A new leap in AI security!
OpenAI says its upcoming model, Astra, can detect previously unknown security vulnerabilities and build chains of exploits without human guidance at every step. The model also discovered two Zero-Day vulnerabilities during internal tests.
⚠️ Most importantly: OpenAI classified Astra as a "Critical Cybersecurity Capability" for the first time, so it will restrict access to its strongest cyber capabilities and add additional layers of protection before launch.
AI no longer just detects bugs… it has started to understand how they can be exploited. 🤖🔐
#OpenAISaysAstraFindsFlawsAutonomously
OpenAI says its upcoming model, Astra, can detect previously unknown security vulnerabilities and build chains of exploits without human guidance at every step. The model also discovered two Zero-Day vulnerabilities during internal tests.
⚠️ Most importantly: OpenAI classified Astra as a "Critical Cybersecurity Capability" for the first time, so it will restrict access to its strongest cyber capabilities and add additional layers of protection before launch.
AI no longer just detects bugs… it has started to understand how they can be exploited. 🤖🔐
#OpenAISaysAstraFindsFlawsAutonomously
