Breaking

OpenAI Says Autonomous Pipeline Found and Patched 948 Linux Kernel Privilege Escalation Flaws in One Run

November 4, 2026 at 18:14 EST

Security
Software Dev & Coding

OpenAI said its security researchers built an autonomous pipeline that discovers, exploits, and patches software vulnerabilities, and that a single run using an upcoming cyber model validated and generated patches for 948 local privilege escalation vulnerabilities in the Linux kernel.

Security research · OpenAI

An autonomous AI pipeline found, exploited, and patched 948 Linux kernel privilege-escalation flaws in one run

The system handles the full vulnerability cycle in a single automated process. The run used an upcoming cyber model that has not been publicly released.

948

Vulnerabilities validated and patched in a single run

Automated run covering discovery, exploit confirmation, and patch generation

Linux

Kernel flaws, in the core layer used by servers, cloud infrastructure, and devices

The pipeline loop

One automated process from flaw to fix

Step 1

Discover

Find candidate flaws

→

Step 2

Exploit

Confirm each flaw can be exploited

→

Step 3

Patch

Generate fixes for validated flaws

↺ The cycle is described as one continuous automated run, not separate manual handoffs.

Scale of the run

Each block ≈ 50 vulnerabilities; 19 blocks ≈ 948

Continue reading

The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.