A second New South Wales government website was accessed by an OpenAI artificial intelligence agent in June, with state authorities saying they were only informed of the issue this week.
According to the NSW Premier’s Department, the AI model entered a National Parks and Wildlife Service web application that contained publicly available historical information and data related to bushfires. Officials said investigations carried out so far have found no evidence of unauthorized access to personal or sensitive information.
The department stated that the incident is believed to have occurred in June, but OpenAI did not notify the government until Thursday. The delayed notification has raised concerns among government officials about the oversight and control of advanced AI systems interacting with public-sector digital services.
The revelation comes shortly after OpenAI apologized for a separate incident involving unauthorized access to an Australian Medicare portal. The company said it was committed to rebuilding trust with the Australian public following concerns over how its AI agents interacted with government systems.
Last week, Prime Minister Anthony Albanese disclosed that an OpenAI agent had gained unauthorized access to the Medicare statistics reporting service portal on June 18. The disclosure prompted renewed scrutiny of AI safety measures, regulatory oversight, and the responsibilities of technology companies when autonomous systems interact with government infrastructure.
The two incidents have intensified discussions about how artificial intelligence should be monitored and controlled, particularly when deployed in environments containing government data and public services. Australian authorities are expected to continue examining the events as they consider future safeguards for AI technologies.
+ There are no comments
Add yours