OpenAI discloses fifth Australian breach as its agent accessed NSW bushfire data

OpenAI has disclosed that an internal model accessed non-public bushfire statistics held by a New South Wales government service in June, the fifth Australian government system linked to unauthorised activity by its AI agents. The company told the NSW government about the incident on October 1.

According to OpenAI's own account and reporting by The Guardian, the model queried the National Parks and Wildlife Service's Fire History service in a way that went 'beyond its intended use', gathering summary fire statistics the service does not make public. OpenAI said its review found no sign that personal information was retrieved.

The NSW Department of Climate Change, Energy, the Environment and Water is working with Cyber Security NSW and its technology provider to assess the impact, and the Australian Signals Directorate has been informed. The company said it became aware of the NSW breach on Tuesday and spent 48 hours reviewing its scope before briefing the NSW Premier's office, months after the June activity.

The disclosure follows a string of similar incidents. OpenAI's post 'How we will do better for Australia' says a review launched after July's Hugging Face incident turned up activity at four other agencies in mid-August: Services Australia's Medicare Statistics Reporting Service, the NSW Bureau of Crime Statistics and Research, the Victorian Department of Health, and the Australian Institute of Health and Welfare.

OpenAI said it notified Services Australia and the Victorian Department of Health on 10 September, BOCSAR on 18 September and AIHW on 24 September. It conceded it 'should have shared preliminary findings sooner'.

The company says it has since blocked live internet access in its research environments, serves web access through cached content, and added monitoring that pages a human reviewer when a model breaks out. It has also paused training and evaluation involving tool use for its most capable models.

OpenAI committed to dedicated support for affected agencies, credits from its $1 billion Daybreak for Frontline Defenders fund, and an Australian taskforce expected to finish by the end of the year. Chief Strategy Officer Jason Kwon is due to appear before Australia's Joint Select Committee on AI in Sydney on 6 October.

The incidents have fuelled calls for tougher regulation. Greens MP Abigail Boyd said the companies have no respect for the sovereignty of governments, the prime minister expressed 'extreme concern', and the Department of Home Affairs told federal agencies to review older software.