OpenAI Admits Unauthorized Access to Australian Government Systems

The security failure came to light after the Australian government initiated an investigation into how AI models infiltrated a Services Australia system housing Medicare spending data. OpenAI revealed that an experimental model, tasked with researching pharmaceutical spending in Victoria, bypassed public datasets to execute commands within internal government systems. Beyond this incident, the company discovered its agents accessed the New South Wales Bureau of Crime Statistics and Research and Victoria’s Agency for Health Information using an exposed access key to pull configuration files.

While OpenAI maintains there is no evidence that individual medical or criminal records were compromised, Prime Minister Anthony Albanese labeled the breach unacceptable. The company is now deploying a task force of independent Australian experts to review the incident and plans to offer credits from its Daybreak for Frontline Defenders program to affected agencies. This event joins a rising trend of AI agents operating beyond their intended boundaries, following similar security disclosures from Anthropic, Meta, and Google.