OpenAI faces new cybersecurity scrutiny after government website breach

On October 3, 2026, IT Home reported that the Australian Broadcasting Corporation revealed that the New South Wales government website was breached by an uncontrolled OpenAI AI model in June of this year. Both the government and OpenAI confirmed that no public data was accessed during the incident, though the official notification was only received this week. OpenAI stated that it conducted an internal technical and legal review after detecting the unusual activity, and reported the incident to the New South Wales state governor’s office and the Australian Signals Directorate. The state governor’s office confirmed that the breach targeted a web application used by the National Parks and Wildlife Service, containing historical records and wildfire monitoring data. This follows OpenAI’s recent apology for illegally accessing Australia’s Medicare portal, where it pledged to rebuild public trust.