{
  "id": 18303731,
  "title": "OpenAI identifies 948 Linux privilege escalation flaws using AI",
  "url": "https://urgent.news/2026/11/05/openai-identifies-948-linux-privilege-escalation-flaws-using-ai",
  "topic": "ai",
  "section": "AI",
  "published": "2026-11-05T07:24:33.000Z",
  "source": {
    "name": "Arabian Post",
    "slug": "arabian-post",
    "url": "https://thearabianpost.com/openai-identifies-948-linux-privilege-escalation-flaws-using-ai/"
  },
  "original_language": "en",
  "account": "OpenAI's research team has uncovered 948 Linux kernel privilege escalation vulnerabilities using an artificial intelligence-powered cybersecurity tool. The findings were revealed during a ten-day investigation, where the experimental pipeline generated workable exploits for the discovered flaws. The investigation consumed $1,213,430 in application programming interface credits, approximately $1,279 per exploit. The vulnerabilities, which allow a user with limited access to gain higher permissions, were tested on virtual machines running a standard Linux installation. OpenAI tested its exploits against the latest Linux kernel development branch but did not disclose the specific distribution or upcoming model. They prioritized the critical vulnerabilities and collaborated with the Linux kernel community to submit additional reports. The automated system selected areas of kernel code for inspection, searched for possible defects, and attempted to reproduce them in isolated environments. The study also discovered additional techniques considered novel, which external kernel exploitation specialists had not encountered before. While the research highlights the potential for AI to uncover critical software weaknesses, it also raises concerns for organizations operating shared Linux infrastructure, as a compromised low-privilege account could lead to wider system control.",
  "summary": "OpenAI has disclosed that an automated cybersecurity system generated working exploits for 948 distinct Linux kernel privilege escalation vulnerabilities during a ten-day research campaign, highlighting the growing ability of artificial intelligence to uncover serious software weaknesses at scale. The company said on Wednesday that its experimental pipeline, powered by an unreleased cyber-focused…",
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 4,
    "also_reported_by": [
      {
        "outlet": "Techmeme",
        "title": "OpenAI details a pipeline that used an upcoming cyber model to validate and patch 948 local privilege escalation vulnerabilities in Linux for $1.2M in API costs (OpenAI)",
        "url": "https://urgent.news/2026/11/05/openai-details-a-pipeline-that-used-an-upcoming-cyber-model-to",
        "published": "2026-11-05T02:50:01.000Z"
      },
      {
        "outlet": "CoinDesk",
        "title": "OpenAI says its AI can break into Linux at scale. Here's what it means for crypto",
        "url": "https://urgent.news/2026/11/05/openai-says-its-ai-can-break-into-linux-at-scale-heres-what-it-means",
        "published": "2026-11-05T08:26:26.000Z"
      },
      {
        "outlet": "Phoronix",
        "title": "OpenAI Finds 948 Local Privilege Escalation Bugs In The Linux Kernel",
        "url": "https://urgent.news/2026/11/05/openai-finds-948-local-privilege-escalation-bugs-in-the-linux-kernel",
        "published": "2026-11-05T10:00:58.000Z"
      }
    ]
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}