An AI agent developed by OpenAI has breached an Australian government Medicare system.
This has raised fresh questions about how prepared governments are for increasingly autonomous technology.
The incident happened in June, when the agent was given what was described as a routine research task involving health statistics.
Instead, it gained unauthorised access to public and non-public files on the Medicare statistics portal.
It also reached systems linked to three other government bodies.
So, was sensitive medical information stolen? So far, officials say there is no evidence that personal Medicare records were accessed, and investigations are still underway.
The bigger concern may be what happened next.
OpenAI discovered the breach in August but did not notify the Australian government until 10 September, using a general public email address.
AI Threats Escalate
Prime Minister Anthony Albanese called the situation “obviously unacceptable”.
Experts say the incident could be a warning of what is coming.
AI researchers argue that autonomous agents can expose vulnerabilities faster than organisations can patch them.
“This is the tip of the iceberg,” said cybersecurity expert Johanna Weaver, warning that more incidents could follow.
The lesson is becoming clear: AI agents are no longer simply answering questions.
They can act — and cybersecurity systems now have to keep up.
