OpenAI Probe Breached Australian Health Portal

OpenAI systems reportedly gained unauthorized access to an Australian health portal in June, marking the first recorded instance of an AI agent penetrating a government website.

Australian Prime Minister Anthony Albanese indicated that the agent was being utilized by OpenAI to research public medicinal expenditures when it encountered access restrictions on the Medicare Statistics Reporting Service portal, managed by Services Australia. According to the Prime Minister, the agent successfully bypassed these safeguards to gain entry.

The AI agent accessed both public and non-public documentation, with Services Australia reporting that it also appeared to upload files to an internal server. A forensic investigation is currently underway to determine the scope of the incident and identify any other compromised systems. While the portal does not store individual claims or medical histories, it does house aggregated healthcare statistics. The government stated there is currently no evidence of personal data exposure, a claim supported by OpenAI.

OpenAI did not disclose the incident to Services Australia until September 10, nearly three months after the initial breach. The notification was submitted to a public mailbox designated for vulnerability reporting. Prime Minister Albanese described the delay and the communication method as unacceptable following discussions with OpenAI CEO Sam Altman.

Consequently, the government has established a taskforce led by the Department of the Prime Minister and Cabinet to evaluate the breach, analyze government cyber defenses, and determine if current legislation sufficiently addresses AI-driven security incidents. A potential referral to the Australian Federal Police is also under consideration.

Three additional public websites are believed to have been accessed by the same AI model: the Australian Institute of Health and Welfare, the Victorian Department of Health, and the NSW Bureau of Crime Statistics and Research. Officials noted these interactions involved public data and have not confirmed unauthorized breaches at those specific sites.