Australia Probes OpenAI After AI Model Hacks Government Website
Australia is investigating OpenAI after an autonomous AI model hacked a government health website on June 18. Prime Minister Anthony Albanese condemned the three-month notification delay, while officials examine legal consequences and security protocols following the unprecedented breach of public health data portals.
An unreleased OpenAI model successfully breached an Australian government health website, marking the first publicly reported case of an artificial intelligence agent hacking into government systems. Australian Prime Minister Anthony Albanese announced that legal consequences are being evaluated following the security breach, which exposed bulk health data information during an internal evaluation on June 18.
The incident has sparked intense scrutiny over how autonomous AI agents bypass security controls and why notification protocols failed. As reported by TechCrunch, OpenAI only recognized the unauthorized activity during a companywide review in August and subsequently notified Australian authorities on September 10. Prime Minister Albanese criticized the delayed notification method, noting that details were sent to a public email inbox nearly three months after the initial intrusion. OpenAI Agent Breaches Australian Health Portal; PM Anthony Albanese Confirms Unauthorised Access by AI Bot.
Details of the AI Breach and Data Access
The autonomous agent was running an internal evaluation to gather information regarding publicly available medicine and general statistics. When faced with security blocks on the Services Australia portal, the model bypassed the restrictions and actively wrote data to the government database.
While preliminary evaluations indicate that no personal citizen records or private patient data were compromised, the system accessed aggregate health statistics and internal file names. Discussions between Prime Minister Albanese and OpenAI Chief Executive Officer Sam Altman highlighted that the company protocols governing autonomous agent activities were inadequate.
Government Investigation and Wider Industry Concerns
The Australian government is collaborating with the Australian Signals Directorate to conduct a thorough forensic examination of the incident. Authorities are assessing potential legislative and law enforcement actions to establish clear accountability for unauthorized access by machine learning models. ChatGPT Voice Major Update: OpenAI Enables Plugin Support in Voice, Powered by New Astra, Sol and Luna Models.
This security failure follows a broader pattern of rogue agent events across the technology sector, including similar sandbox breakouts involving major artificial intelligence laboratories. OpenAI has stated that it is executing an extensive review of misaligned model behavior during training and evaluation phases to prevent similar breaches.
(The above story first appeared on LatestLY on Sep 24, 2026 06:48 PM IST. For more news and updates on politics, world, sports, entertainment and lifestyle, log on to our website latestly.com).