Security · High impactFan score 46 · Tier B3 sources353 comments in the community
AI Agents Got Root on Linux via 948 Kernel Flaws in OpenAI Research
SecurityWeek · By Eduard Kovacs ·
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 .
The Brief
- OpenAI built an autonomous pipeline that found, validated, and generated patches for 948 local privilege-escalation vulnerabilities in the Linux kernel.
- SecurityWeek reports that an unreleased OpenAI cyber model found and exploited the 948 bugs at an API cost of about $1.2 million.
- OpenAI says the post omits full technical details of the new exploitation techniques for safety reasons.
Why it matters
A local privilege escalation lets someone with a low-privileged account on an unpatched Linux system gain root, so systems that are not up to date are at real risk and need the kernel fixes applied.
Who’s affected
- Teams running Linux kernels on servers, workstations, or embedded devices
- Operators of systems where users have local accounts
What to do
- Apply Linux kernel updates that include these fixes as your distribution ships them
- #linux
- #kernel
- #security
- #privilege-escalation
- #ai-security
- #openai
What developers are saying
Based on 353 comments · Mixed reactions
Commenters on Hacker News are split. Some see cheap AI-driven scanning as a way to improve security for open-source projects, while others doubt the claims because the post attaches little data and the bugs may overlap or already be public. Several also note that many IoT devices will never receive these updates and argue that deeper architectural change is needed.
- Skepticism about marketing claims25%
- Kernel design and memory-safe languages25%
- AI as a tool for open-source security20%
- Unpatched IoT devices and real-world risk15%
- Cost compared with traditional tools10%
In favour
- + Cheap AI-driven scanning could give open-source projects security auditing they cannot otherwise afford
- + Repeated scans with each new frontier model could help eliminate latent bugs in large codebases
Concerns
- − The post has no real data attached and many of the 948 bugs may be variants of the same few issues
- − Patches were reportedly already public, so the withheld technical details are hard to justify
- − Thousands of IoT devices will not receive these kernel updates
- − Monolithic kernels need a better architecture, such as a microkernel, to solve the problem
“Imagine if open-source organizations other than the Linux Foundation (and Mozilla I guess) could throw $1.2M of API-rate LLM usage at security.”
“There is no real data attached to the post.”
“and yet thousands of iot devices will not get these updates. This is gonna be fun.”
Timeline
Mid-October scan of the latest Linux kernel master branch finds 948 unique LPE vulnerabilities
OpenAI publishes the results of the pipeline
OpenAI News
A thousand fewer paths to root
Hacker News: Front Page
A thousand fewer paths to root
Hacker News
479 points · 353 comments
SecurityWeek reports on the 948 kernel flaws
SecurityWeek
AI Agents Got Root on Linux via 948 Kernel Flaws in OpenAI Research
Full coverage · 3 sources
- 18h ago ↗
SecurityWeek
AI Agents Got Root on Linux via 948 Kernel Flaws in OpenAI Research
- 2d ago ↗
Hacker News: Front Page
A thousand fewer paths to root - 2d ago ↗
OpenAI News
A thousand fewer paths to root
