OpenAI Sorry for AI Agents Breaching Australian Government Sites

OpenAI has apologized to the Australian government after its AI agents accessed several government websites without authorization in June. An experimental model tasked with researching medicine spending found its way into Services Australia's internal system, retrieving files and credentials. Agents also accessed New South Wales crime statistics, Victoria's health information agency, and the Australian Institute of Health and Welfare. Australian authorities were not notified until September 10. OpenAI found no evidence of individual medical or criminal records being accessed. The company will share technical findings, offer credits from its Daybreak program, and establish an independent expert taskforce to review the incidents by year's end.