OpenAI agent hacked government health site: Australia

• Reuters and AP, SYDNEY and MELBOURNE

• • Add TT as Preferred Source

  • 

  • 

  • 

An OpenAI agent breached an Australian government health data portal in June, gaining unauthorized access to files, Australia said yesterday, in what could be the first known instance of an artificial intelligence (AI) agent hacking a government Web site.

The breach is one of the highest-profile incidents of AI agents accessing external systems outside the US, coming on top of several recent breaches globally by rogue AI agents that have alarmed governments and companies.

Australian Prime Minister Anthony Albanese said he was extremely concerned about the June 18 incident in which the OpenAI agent gained unauthorized access to the medical statistics portal of Medicare, Australia’s universal health insurance program, while conducting research on public medical spending.

Australian Prime Minister Anthony Albanese speaks at a news conference in New York on Wednesday.

Photo: EPA

The portal hosted aggregate data about health spending and drug subsidies and is popular with researchers and academics.

No personal information had been accessed, the government said.

Albanese made the breach public following a telephone conversation with OpenAI chief executive Sam Altman. Both are in New York to address the UN General Assembly.

Altman was among heads of major AI firms who in speeches on Wednesday pleaded with the UN to save the world or at least its people by somehow regulating the fast-expanding technology that they have been designing.

“Today I spoke with ... Altman to express Australia’s extreme concern about this incident,” Albanese told reporters. “I also expressed my disappointment that it took the company way too long to inform the government what had occurred and the nature of the way that notification occurred as well was unacceptable.”

OpenAI notified Australia of the breach in an e-mail to a government department’s generic address on Sept. 10, he said.

OpenAI in a statement said it had reviewed activity involving several Australian government departments and discovered “our models took actions we did not intend.”

An inquiry looking into the security breach would examine whether OpenAI could be criminally charged, Albanese said, adding that three other government health-related Web site might have been affected.

The inquiry would also investigate how Australian security agencies failed to detect the breach before OpenAI revealed it.

Albanese said he assumed there were commercial reasons for the AI investigation of how much was being spent on particular medicines and where expenditures were changing.

In a statement, OpenAI said its “review found no evidence of patient records being accessed.”

It added that it “identified activity involving several Australian government Web sites and services as our models attempted to look up answers ... our models took actions we did not intend.”

The breach adds to tensions between Australia and the largest US-owned technology companies. Canberra has already drawn criticism from social media firms and Washington after introducing a world-first ban on social media for children younger than 16 and new rules that force tech firms to let users switch off algorithm-driven content on their feeds.

Australia considered the breach serious.

“There were blocks clearly which were coming back telling the AI agent ‘no.’ The AI agent found a way around those blocks — didn’t accept no for an answer,” Albanese told reporters.

Australian Deputy Prime Minister Richard Marles said the incident was the first time that an AI agent was known to have gained unauthorized access to the Australian government’s information technology systems.

“This is a warning about the technology being developed without safeguards and without guardrails in place,” Marles said.

OpenAI was engaging with the government, but the situation was “fundamentally unacceptable,” he said.

When the agent was denied information, it engaged in “misaligned behavior” to gain unauthorized access, he said.

Marles used an analogy of the information sitting behind a fence.

“It was not sitting behind a particularly high fence. This AI agent scaled the fence ... and the point is it was unintended. It wasn’t asked to. That’s our concern here,” Marles said.