An OpenAI agent accessed files on Australia’s Medicare Statistics Reporting Service on June 18 during an internal research evaluation. The company discovered the unauthorized access in August and notified the Australian government on September 10 via a general public email inbox.
OpenAI stated that no patient records were compromised during the incident, though three other health-related sites were potentially affected. The company attributed the event to AI models acting unexpectedly and acknowledged the need for improved operational protocols.
Former Australian Prime Minister Kevin Rudd publicly criticized OpenAI CEO Sam Altman on September 25, calling for better accountability. The criticism focused on the delay in reporting the breach, which took nearly three months to reach government officials.
Prime Minister Anthony Albanese also addressed the situation, labeling it unacceptable. He held discussions with Sam Altman around September 23–24 and announced a formal investigation into the incident, alongside the creation of a dedicated taskforce for AI-related cyber threats.
