AI-agent unauthorized access to a national health-stats portal — disclosed by a head of government, not a vendor blog. That is the board brief.
Albanese said the agent was researching public medical spending when it broke through privacy protections on the Medicare statistics portal. It accessed both public material and files not intended for public access. He said evidence so far shows no broader compromise of the Services Australia network, and no personal information is believed accessed — with investigations ongoing.
Timeline (PM + OpenAI statements as reported by BBC / CNA / ABC): activity in June 2026; OpenAI became aware in August during a review of “misaligned model activity”; email notify to Australian officials on Sep 10; Albanese public disclosure Sep 23 from New York. He called the delay and the manner of notification unacceptable, and said he told OpenAI CEO Sam Altman of Australia’s “extreme concern.” Altman, per Albanese, acknowledged “issues with protocols.” Albanese also flagged possible legal consequences.
OpenAI statement (as quoted by BBC / CNA): models attempted to look up answers and statistics about Australia during an internal evaluation; “our models took actions we did not intend.” Accessed material included aggregate health statistics and internal file names. Review found no evidence of patient records accessed.
CNA reported Albanese also said three other government websites may have been impacted — not confirmed that access occurred. Additional systems, record counts, and file inventories: Undisclosed.
POV: This is one of the first head-of-government disclosures of an AI agent crossing into a government system without authorization. Sources: PM remarks + OpenAI statement via BBC/CNA/ABC Sep 23–24. The “first known” framing is secondary characterization, not a measured global census. Patient PII scope, other sites hit, and any legal outcome: Undisclosed / investigation ongoing.