OpenAI says AI pipeline produced 948 Linux privilege escalation exploits

First reported by RuntimeWire at · Updated · 7 sources

The prototype used an upcoming cyber model during a mid-October research run and generated candidate patches for maintainer review. The vulnerabilities could let an unprivileged local user gain root access. According to The Hacker News, the run lasted 30 hours and cost $1.2 million in API usage.

  • The pipeline reproduced 1,724 kernel vulnerabilities.
  • RuntimeWire reports it also produced 291 information leak proofs of concept.

Covered by 7 publishers within 11 hours of the first report.

Reporting7

RuntimeWire OpenAI says its agents exploited and patched 948 Linux kernel flaws for $1.2M · Ryan Merket
Unite.AI OpenAI Pipeline Finds 948 Linux Kernel Privilege Escalation Flaws · Miles Okada, AI & Cybersecurity, AI Research Agent
Cyber Security News OpenAI Cyber Model Uncovers 948 Linux Kernel Privilege Escalation Vulnerabilities · Guru Baran
The Hacker News OpenAI AI Pipeline Generates 948 Linux Root Escalation Exploits in 30-Hour Run · info@thehackernews.com (The Hacker News)
Neowin OpenAI's upcoming model finds 948 Linux kernel vulnerabilities that could grant root access · Pradeep Viswanathan
CoinDesk OpenAI says its AI can break into Linux at scale. Here's what it means for crypto · Omkar Godbole
Phoronix OpenAI Finds 948 Local Privilege Escalation Bugs In The Linux Kernel · Michael Larabel

Related

Xint shows OpenAI compaction flaw can expose hidden AI reasoning
OpenAI staff raised security concerns with leaders for years
Haiou jailbreak alters OpenAI Codex instructions to insert ads
Irregular blames human error for AI agents escaping tests

Topics OpenAI