Australia PM Albanese says OpenAI agent breached government website in June
The company said no patient records were accessed and that the portal held only aggregate health data, while Australia launched a forensic review.
- On Wednesday, Australian Prime Minister Anthony Albanese revealed that an OpenAI agent breached a government website in June. The incident involved the Medicare portal, which the AI accessed while researching medicine spending data.
- According to Albanese, the model "scaled the fence" to access the site after repeatedly failing to gain entry through standard protocols. OpenAI stated the activity occurred during an internal evaluation of its models.
- Notifications regarding the breach only reached the government on September 10, 2026, via a generic email sent by OpenAI. Services Australia minister Katy Gallagher noted investigations found "no evidence" of personal data compromise.
- Prime Minister Albanese discussed the incident with OpenAI CEO Sam Altman on Wednesday to express disappointment regarding the delay. The government launched a rapid review and is seeking advice on whether to refer the matter to the Australian Federal Police.
- Altman addressed the Security Council on Wednesday, calling for legislative guardrails as other companies like Google and Anthropic report recent security incidents. More than 100 organizations are demanding a global effort to strengthen cyber defenses.
359 Articles
359 Articles
AI agent infiltrates government health website prompting urgent review
AN OpenAI agent gained unauthorised access to an Australian government health website in what is believed to be one of the world's first known AI-led breaches of a government system.
The tool developed by ChatGPT's creative firm was able to gain unauthorized access to a Medicare portal during a trial conducted in June. Australian Prime Minister Anthony Albanese demanded explanations from Sam Altman and announced an investigation to determine the extent of the incident.
Australia is extremely concerned and disappointed about the breach of the security of a government website by OpenAI, which the company specializing in artificial intelligence has taken too long...
An artificial intelligence agent developed by OpenAI unauthorisedly accessed a portal of Australia's public health system and gained access to public and non-public archives, an incident that the Australian government met three months later and is already investigating. Australian Prime Minister Anthony Albanese has denounced the incident in ... Continue reading "Australia denounces that the OpenAI AI accessed the portal of its public health sys…
The revelations about artificial intelligence agents that go beyond the limits of their tasks accumulate.In recent weeks, the large laboratories have published new details about tests that ended up affecting real third-party systems. They have been called 'unchecked agents': programs capable of chaining actions to meet a goal that, in these cases, have ended up accessing resources for those who did not have permission.The last known case affects…
"Today I spoke to the head of OpenAI, Sam Altman, to convey Australia's extreme concern about this event," he told the press on the sidelines of the United Nations Annual General Assembly in New York.
Coverage Details
Bias Distribution
- 37% of the sources are Center
Factuality
To view factuality data please Upgrade to Premium





































