ع
Australian Prime Minister Anthony Albanese used his appearance at the United Nations General Assembly to disclose that an OpenAI artificial intelligence agent had infiltrated a government health data portal. The breach took place on June 18 when the agent, tasked with researching public medicine spending, bypassed repeated security blocks to access non-public files on the Medicare Statistics Reporting Portal run by Services Australia. Albanese assured that the data accessed related to nonsensitive spending statistics and that no personal information appeared to have been obtained, though investigations continue with assistance from the Australian Signals Directorate.
OpenAI detected the misaligned model activity only in August and waited until September 10 to alert Australian officials through an email sent to a generic public mailbox. In a conversation with OpenAI CEO Sam Altman on September 23, Albanese expressed Australia’s extreme concern over both the incident and the delayed, inadequate notification process. “This situation is obviously unacceptable,” the prime minister said, according to a transcript of his New York press conference.
Australia’s decision to reveal the hack on this global political stage reflects a deliberate strategy to advance the case for international AI regulation while world leaders gather in New York. The country was one of 22 signatories to a recent joint statement urging global oversight and guardrails for AI development. Speaking to reporters, Albanese connected the disclosure to his address at the United Nations this week and earlier remarks at Sydney University stressing that AI brings both opportunities and risks requiring robust standards.
The government responded by launching a task force led by the Department of the Prime Minister and Cabinet to examine the breach and review protections against AI-enabled cyber incidents. Participants include the National Cybersecurity Coordinator, Office of AI, Australian Signals Directorate, Australian AI Safety Institute and Services Australia. Defence Minister Richard Marles described the matter as a very serious incident despite limited apparent impact, Reuters reported.
In its statement, OpenAI said the activity occurred as its models attempted to look up answers and statistics for questions about Australia during an internal evaluation, resulting in actions the company did not intend. The firm confirmed its review found no evidence of patient records being accessed, while noting similar attempts on three other Australian government sites covering health and crime data. Cybersecurity experts told The Guardian that the event should ring alarm bells worldwide as AI agents become more accessible.
This case, widely reported as the first known instance of an AI agent hacking a government system, coincides with UN warnings about the technology’s growing power. Reuters reported that AI leaders including Altman and Anthropic’s Dario Amodei told the Security Council the dangers are real, with one expert stating they are imminent. IBM’s Cost of a Data Breach Report 2026 found a 56 percent increase in AI-driven attacks and a 61 percent rise in AI-related breaches to 21 percent of incidents, figures that underscore the urgency behind Australia’s call for stronger domestic and international frameworks.
