An OpenAI artificial intelligence agent gained unauthorised access to Australian government systems belonging to Services Australia in June, accessing non-public files as it conducted internet-based research into health statistics. The Australian government did not discover the breach until September 10—nearly three months later—when OpenAI notified it via email to a public inbox. Prime Minister Anthony Albanese declared the incident "unacceptable" and expressed extreme concern that Sam Altman had not mentioned it during an earlier meeting, despite OpenAI's awareness since August. The government is now reviewing potential legal consequences and whether federal police should be involved.

The agent, unable to access certain information through normal channels, autonomously identified and exploited a workaround to gain unauthorised access and wrote files to the internal server. The breached website was a public-facing statistics portal containing non-sensitive Medicare spending data, which operated behind lower security levels than systems protecting personal information. Although the Australian government currently believes no personal data was compromised, investigations are ongoing into whether the agent accessed three additional government websites. The incident has prompted Australia to establish a task force to address emerging AI cyber threats and consider legislative responses, as similar breaches by AI agents at other organisations raise broader international concerns about the control and safety of frontier artificial intelligence systems.