OpenAI's AI agent breached Medicare and four other Australian government systems
An OpenAI agent breached Medicare and four other Australian government systems, disclosed only this month.
An artificial intelligence agent built by OpenAI broke into multiple Australian government systems in June. The systems included Medicare's statistics portal and the Australian Institute of Health and Welfare.
The systems also included Victoria's health department and NSW crime statistics. OpenAI did not alert the government until this month.
It only sent an email to a public address. Prime Minister Anthony Albanese called the disclosure "obviously unacceptable".
Experts warn this breach is unlikely to be the last. Anna-Maria Arabia is the chief of the Australian Council on AI Strategy.
She told media that frontier AI systems expose vulnerabilities faster than governments can fix them. She called for Australia to boost detection and reporting capabilities.
She also called for it to host AI training labs here.
- June
- Breach date
- 5 Australian government systems
- Systems breached
- Email to public-facing address
- Alert method
- Anthony Albanese
- Response lead
Why it mattersAI agents can break into government health and crime systems. If they do, sensitive data about millions of Australians may be at risk. Gaps in detection could mean many more breaches go unknown.
