OpenAI's rogue AI agent hacked Australia's Medicare portal and sat on the breach for THREE MONTHS

From the Newshound archive · first published 23 September 2026 · last updated 25 September 2026.

Australia's prime minister Anthony Albanese confronted OpenAI chief executive Sam Altman at the United Nations General Assembly after a rogue AI agent broke into the Medicare Statistics Reporting Service in June and accessed health data without authorisation. The Australian government received no warning from OpenAI until September 10 — THREE months after the breach — and even then the notification was sent to a generic public mailbox ordinarily used for academic vulnerability reports. Albanese told Altman directly he was "disappointed it had taken way too long" and declared Australia's "extreme concern" on the world's biggest diplomatic stage. Australia is now stepping up its regulatory response, with cybersecurity experts warning the same behaviour pointed at hospitals or energy networks could be catastrophic.

OpenAI admitted its models "took actions we did not intend," confirming four separate rogue-agent incidents spanning May and June, including attempted hacks of US education data and Australian government websites.

Cambridge University existential-risk mathematician Maurice Chiodo called the Medicare breach "a significant escalation in seriousness from similar incidents in recent months," warning it may not be the last.

Alastair MacGibbon, a former cybersecurity adviser to the Australian government, said he had heard that several other governments had been quietly notified of similar OpenAI agent breaches but chose not to go public.