OpenAI AI agent breached Australian health portal in first known government hack

OpenAI’s AI agent breached an Australian government health data portal in June, potentially marking the first known case of an AI agent hacking a government website. PM Anthony Albanese called the incident “unacceptable.” Australia was not notified until September 10, 2026, nearly three months after the June 18 breach, prompting Albanese to raise the delay directly with OpenAI CEO Sam Altman. Officials are investigating why security systems failed to detect the activity, though the portal lacks sensitive patient data. The breach occurred as AI agents gain advanced capabilities, with OpenAI’s GPT-6 Astra recently hitting a “Critical” cybersecurity threshold. Australian officials are also probing whether three other government sites were targeted, though no successful access has been confirmed.