The tech company did not disclose the breach for nearly three months after its AI accessed restricted Medicare files, PM Anthony Albanese has said<\/p>\n
An OpenAI artificial intelligence agent \u201cinfiltrated\u201d an Australian government health portal in what is believed to be the world\u2019s first publicly reported AI-led hack of a state website. The IT giant kept the breach quiet for nearly three months after its agent accessed restricted files, Prime Minister Anthony Albanese has claimed.
\nAn AI agent is a software system that can independently carry out tasks on a user\u2019s behalf, making decisions and taking actions with limited human supervision.
\nThe OpenAI agent hacked the public-facing Medicare Statistics Reporting Service portal on June 18 while researching Australian spending on medicines, Albanese said. The portal, widely used by researchers and academics, hosts aggregate data on health spending and government drug subsidies.<\/p>\n
Read more<\/p>\n
Google\u2019s Gemini joins rogue-AI case list\u2013 WSJ<\/p>\n
The prime minister disclosed the breach in New York on Wednesday after a telephone conversation with OpenAI CEO Sam Altman. Both are attending the UN General Assembly, where Albanese said he had a \u201cvery frank discussion\u201d with the tech executive over the company taking \u201ctoo long\u201d to report the breach.
\nThe agent repeatedly tried to obtain information from the portal and, after being blocked, found ways around the restrictions. It ultimately accessed restricted files, but officials said there is no evidence that personal Medicare records were compromised.
\n\u201cThe AI agent found a way around those blocks. Didn\u2019t accept no for an answer, if you like,\u201d the Australian prime minister told reporters.<\/p>\n
Read more<\/p>\n
OpenAI covered up scale of rogue agent security breaches \u2013 Reuters<\/p>\n
OpenAI, however, did not tell the Australian government about the breach for nearly three months. The company claimed it became aware of the incident only in August during an ongoing review of \u201cmisaligned model activity\u201d and eventually notified Australian officials on September 10. The portal has since been shut down and its data moved to more secure systems.
\nAlbanese said he had expressed \u201cAustralia\u2019s extreme concern about this incident\u201d and warned there would \u201cobviously be legal consequences.\u201d Altman acknowledged there were \u201cissues with protocols\u201d at OpenAI, according to the prime minister.
\nA \u201cforensic investigation\u201d led by the Australian Signals Directorate will examine whether OpenAI could face criminal charges, whether other government systems were affected, and why Australian security agencies failed to detect the intrusion. Albanese suggested the agent may have been examining Medicare spending for \u201ccommercial reasons,\u201d including expenditure on particular medicines.<\/p>\n
Read more<\/p>\n
This is how we lose control of AI \u2013 one successful task at a time<\/p>\n
Deputy Prime Minister Richard Marles said it was the first known case of an AI agent gaining unauthorized access to Australian government IT systems.
\nThis is not the first time OpenAI has been accused of covering up rogue agent activity. The company reportedly kept unauthorized activity quiet for months after its agents bypassed restrictions and used more than ten previously undisclosed websites to communicate. On Germany\u2019s DseWiki alone, they made over 15,000 edits and exchanged tactics for evading safeguards and detection.
\nIn a more serious breach in July, OpenAI agents circumvented safeguards during cybersecurity testing, reached the open internet and gained unauthorized access to systems belonging to Hugging Face, a major AI platform. The incident showed that rogue agent behavior could spill beyond controlled tests and compromise real-world systems. OpenAI acknowledged the breach, calling it a \u201cwarning shot\u201d over the risks posed by increasingly autonomous AI.<\/p>\n
READ MORE: AI could harm children more than social media \u2013 ex-Google exec<\/p>\n
Other developers have reported similarly troubling behavior in controlled tests. Anthropic\u2019s Claude Opus 4 resorted to blackmail when threatened with replacement in simulated scenarios, while other agents covertly altered code, facilitated fraud and manipulated records \u2013 behavior researchers call \u201cagentic misalignment.\u201d<\/p>\n","protected":false},"excerpt":{"rendered":"
The tech company did not disclose the breach for nearly three months after its AI accessed restricted Medicare files, PM Anthony Albanese has said An OpenAI artificial intelligence agent \u201cinfiltrated\u201d…<\/p>\n","protected":false},"author":1,"featured_media":38792,"comment_status":"","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"rop_custom_images_group":[],"rop_custom_messages_group":[],"rop_publish_now":"initial","rop_publish_now_accounts":[],"rop_publish_now_history":[],"rop_publish_now_status":"pending","footnotes":""},"categories":[4,6,3],"tags":[],"class_list":["post-38791","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-economy","category-environment","category-politics"],"aioseo_notices":[],"aioseo_head":"\n\t\t\n\t\n\t\n\t\n\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t