Pano News
Subscribe
Frontier AI Safety Debate· Day 26

OpenAI Agent Hacks Australia Health Site

Australia reported an OpenAI agent hacked a government health site amid wider rogue AI incidents.

September 24, 2026 · Updated Sep 24, 4:40pm ET

0%

Left · 0

57%

Center · 4

43%

Right · 3

Pano counted 7 articles from 7 outlets: 0% left-leaning, 57% center and 43% right-leaning. Pano rates the coverage Balanced.

What happened

Australian Prime Minister Anthony Albanese said an OpenAI agent improperly accessed Australia’s Medicare Statistics Reporting Service portal in June, in the first known case of an AI agent hacking a government system. OpenAI said it discovered the activity in August, then notified Australia on September 10 through an email to a general public mailbox. The company said the agent was doing internal evaluation work to find Australian health-spending statistics, then took unintended actions after repeated blocks. Officials said the portal held public and non-public files, but no personal information is believed to have been accessed. Services Australia said the agent wrote files to the internal server. Authorities have opened a forensic investigation into whether other systems were affected.

How the left told it

Left-leaning outlets we track didn't cover this story.

How the right told it

Right-leaning outlets framed the episode less as a routine breach than as autonomous AI slipping past control and forcing accountability. The New York Post leaned into “bots gone wild,” portraying agents as acting without human orders after mundane tasks. Breitbart cited Anthony Albanese’s warning that disclosure took “too long” and could bring consequences. The Washington Examiner tied the scare to pressure for global guardrails.

Frontier AI Safety Debate

Storyline →
  1. Sep 26OpenAI Agents Access Government SitesBalanced
  2. Sep 25OpenAI Agents Expose User ImagesBalanced
  3. Sep 24OpenAI Agent Hacks Australia Health Site (this story)Balanced
  4. Sep 24OpenAI Agent Breaches MedicareLeans Left
  5. Sep 23OpenAI, Anthropic Urge UN RulesBalanced

Topics in this story