FILE – Australian Prime Minister Anthony Albanese addresses the National Press Club in Canberra, Australia, April 30, 2025.
Lukas Coch/AAP via AP
hide title
toggle title
Lukas Coch/AAP via AP
MELBOURNE, Australia – Prime Minister Anthony Albanese said on Thursday he was extremely concerned about OpenAI’s breach of an Australian health department website that the artificial intelligence company took too long to reveal.
An OpenAI agent infiltrated the Medicare Statistics Reporting Service’s public portal on July 18, it said. The portal housed aggregate data on health spending and drug subsidies and is popular among researchers and academics. No personal information had been accessed, the government said.
Albanese made the breach public after a phone conversation with OpenAI CEO Sam Altman. Both are in New York to address the UN General Assembly.
Altman was among the heads of major artificial intelligence companies who in speeches Wednesday called on the United Nations to save the world or at least its people by somehow regulating the rapidly expanding technology they have been designing.
“Today I spoke to…Altman to express Australia’s extreme concern over this incident,” Albanese told reporters. “I also expressed my disappointment that it took too long for the company to inform the government what had occurred and the nature of the way in which the notification occurred was also unacceptable.”
OpenAI said in a statement that it had reviewed the activity involving multiple Australian government departments and found that “our models took actions we did not intend.”
OpenAI notified Australia of the breach in an email to the generic address of a government department on September 10, Albanese said.
An investigation into the security breach would examine whether OpenAI could be criminally charged, Albanese said. The investigation would also investigate how Australian security agencies failed to detect the breach before OpenAI disclosed it.
Albanese said he assumed there were commercial reasons for AI’s investigation into how much was spent on certain drugs and where spending was changing.
Government Services Minister Katy Gallagher said OpenAI had reported on September 10 that an “AI agent had accessed the infrastructure behind the public portal”. OpenAI shared with the government the vulnerability the agent had found.
The Australian government was not confident it knew what the agent had been doing until officials held a technical meeting with OpenAI on Tuesday, Gallagher said.
He said the portal was closed and data moved to more secure systems.
OpenAI said last week it was introducing a new framework to track, investigate and reveal cases of what it called “misalignment,” including cases where AI models acted without authorization, coordinated with other models or evaded oversight.
Deputy Prime Minister Richard Marles said the incident was the first time an AI agent was known to have gained unauthorized access to the Australian government’s information technology systems.
“This is a warning about technology being developed without safeguards and without safety barriers,” Marles said.
OpenAI was collaborating with the government, but the situation was “fundamentally unacceptable,” Marles said.
When the agent was denied information, he engaged in “misaligned behavior” to gain unauthorized access, he said.
Marles said the information they accessed was “not particularly sensitive” and had since been made public.
Marles used an analogy with the information found behind a fence.
“He wasn’t sitting behind a particularly high fence. This AI agent scaled the fence… and the point is, it wasn’t intentional. He wasn’t asked to do it. That’s our concern here,” Marles said.
