An unreleased OpenAI cyber model found and exploited 948 distinct Linux kernel privilege escalation bugs for about $1.2 million. The post AI Agents Got Root on Linux via 948 Kernel Flaws in OpenAI Research appeared first on SecurityWeek.
این صفحه فقط عنوان و خلاصه کوتاه خبر را نمایش می‌دهد. متن کامل در منبع اصلی منتشر شده است.
شرکت OpenAI از مدل هوش مصنوعی جدیدی به نام GPT-۶ Cyber رونمایی کرده که به‌طور تخصصی برای پیدا کردن حفره‌های امنیتی و بررسی امکان سوءاستفاده از آن‌ها ساخته شده است. توانایی‌های این مدل به‌اندازه‌ای پیشرفته ارزیابی شده که OpenAI آن را در بالاترین سطح ریسک سایبری خود با درجه «Critical» قرار داده است.
Decade-old vulnerability in OpenSSH is now fixed, so get a-patchin'
A critical authentication bypass, with no CVE assigned yet, was disclosed in OpenSSH on musl-based Linux distributions such as Alpine. An unauthenticated attacker can log in over SSH password authentication as an account that should not accept passwords at all, including root. On a default Alpine install that means root access from a single connection, […]
OpenAI has developed an autonomous AI pipeline that can identify software vulnerabilities, develop working exploits to validate their impact and generate patches for human review. In an October experiment targeting the Linux kernel, the pipeline created exploits for 948 distinct local privilege-escalation (LPE) vulnerabilities at a total API cost of $1.2 million, the company said in a November 4 research report, “A t…
