OpenAI says its upcoming Astra model can autonomously discover previously unknown software vulnerabilities and develop working exploits with minimal human guidance.
ℹ️ In testing, Astra scored 100% on a benchmark for developing exploits from known vulnerabilities and reportedly found two previously unknown flaws during an internal test.
Because of these capabilities, OpenAI has classified Astra as having “Critical” cybersecurity capabilities and is adding stronger safeguards before wider access.
⚠️Why it matters: For crypto and other online systems, AI that can find and exploit vulnerabilities much faster could make cyberattacks significantly more dangerous.
AI is getting scary good at finding bugs… hopefully the hackers don’t get the premium subscription 🙂
Don’t forget to join our giveaway⭐
🙂 @tele24hr | 🙂 @tele24hrchat