AI··1 min readTechnical readSource-backed

PHP armageddon: OpenAI, Google’s Wiz find ancient holes, panicking the internet

The U.S. National Security Agency warned that two critical security holes have been discovered in PHP, a foundational server-side programming language that powers roughly 70 percent of internet websites.

By
WIREUNWIRED BRIEF

The U.S. National Security Agency warned that two critical security holes have been discovered in PHP, a foundational server-side programming language that powers roughly 70 percent of internet websites. The PHP Foundation has already issued patches for the vulnerabilities, which affect all versions of the software released since 2004, including popular content management systems like WordPress.

OpenAI and Google’s Wiz disclosed that they independently uncovered the decades-old flaws using artificial intelligence tools. Security testing firm Calif used OpenAI's GPT-5.6 Sol under a "Patch the Planet" initiative, while the Wiz team utilized its Wiz Atlas security agent alongside Gemini 4 Argon and GPT-5.6 Sol. Both organizations followed responsible disclosure practices before publicizing the issues.

The vulnerabilities allow attackers to infiltrate underlying server file systems or connected databases. While no exploits have been reported in the wild prior to disclosure, Wiz warned that large language models remove previous barriers by helping attackers adapt exploits across different PHP versions and memory layouts. The NSA advised that websites unable to be patched immediately should be taken offline until updates are complete.

Because versions 8.1 and older are no longer supported, administrators must update their web servers to the current version, 8.5.12.

WordPress stated the

STORY JOURNEY

The story so far

Earlier coverage connected by shared names and topics. Follow the dates to see how coverage developed; these links do not imply every article describes the same event.

  1. Editorial illustration for OpenAI, Google pushing new AI models easily tricked into making nudes, researchers say
    OpenAI, Google pushing new AI models easily tricked into making nudes, researchers say

    Safety researchers have exposed significant flaws in the guardrails of leading AI image generators from OpenAI and Google. According to a preprint…

    Connected by: OpenAIEARLIER COVERAGE
  2. Editorial illustration for FBI Arrests OpenAI Engineer Accused of Hiding Recording Device in Restricted Conference Room
    FBI Arrests OpenAI Engineer Accused of Hiding Recording Device in Restricted Conference Room

    Federal authorities have arrested an OpenAI engineer accused of hiding a recording device behind a television in a restricted executive conference room.…

    Connected by: SecurityEARLIER COVERAGE
  3. Hackers hit AI coding tool OpenCode, running up charges on users’ AI accounts

    Attackers breached an older system belonging to open-source AI coding agent OpenCode, compromising digital passwords and running up hundreds of dollars in…

    Connected by: OpenAIEARLIER COVERAGE
  4. Google is bringing a more personal version of its Gemini AI assistant to your car

    Google is updating its vehicle-based AI assistant to pull data from a driver’s personal Google apps and services. The upgrade aims to…

    Connected by: GoogleEARLIER COVERAGE
  5. NOW
    PHP armageddon: OpenAI, Google’s Wiz find ancient holes, panicking the internetCURRENT UPDATE
A WireUnwired summary is already available. You can still write your own brief below for the reader community.
SOURCE & ATTRIBUTION
TechCrunch

Original author: Julie Bort

Original publication: 21 October 2026 23:29

Read full story on TechCrunch ↗
RELATED STORIES
Editorial illustration for OpenAI Quietly Snaps Up Defunct AI Travel Agent OdessiaTechCrunchOpenAI Quietly Snaps Up Defunct AI Travel Agent OdessiaTechCrunchPalo Alto Networks targets AI agent security with potential acquisitionTechCrunchCrowdStrike Bets Nearly Half a Billion on the Next Security Frontier
Original sourceTechCrunch ↗