An OpenAI artificial intelligence (AI) model gained unauthorized access to at least one Australian government website, Prime Minister Anthony Albanese said on Wednesday.
Albanese said an OpenAI agent accessed public and non-public files on the Medicare statistics portal. Medicare is Australia's publicly funded universal health insurance scheme. Three other government systems may also have been impacted, Albanese added.
No personal information is believed to have been accessed, but investigations were ongoing, Albanese told reporters on the sidelines of the UN General Assembly in New York.
"Nonetheless, this situation is obviously unacceptable," Albanese said.
"I think OpenAI know that they need to have better protocols in place. And they're one of the businesses that themselves have warned of the risks which are there," Albanese said.
The incident occurred in June, but Australia was not notified until September via an email sent to a public mailbox, Albanese said.
Albanese said he had spoken with OpenAI chief executive Sam Altman to express Australia's "extreme concern" about the incident and the way the government had been notified.
"It was a shock that it occurred, because it was real and serious. But it also, I think, was something that had been predicted, including by the AI companies themselves have said that one of the risks that we need to deal with here is that artificial intelligence can go its own way."
OpenAI said in a statement the breach was discovered in August during a review.
"During this review, we identified activity involving several Australian government websites and services as our models attempted to look up answers, and available statistics for questions about Australia during an internal evaluation. In the course of that, our models took actions we did not intend."
The company said it had found no evidence of patient records being accessed.
"We notified the organizations and are providing technical information to support their investigations and help address potential security vulnerabilities. Our overall review is ongoing, and we remain committed to transparency about these issues and to sharing what we learn as that work continues."
Albanese said the incident was a "clear illustration" of why the country was moving to establish standards for AI.
"Standards to ensure that AI will work and deliver for people, not the other way around."
News of the breach came as AI leaders, including Altman, appealed for international cooperation on the technology at the UN Security Council.
The incident follows a series of high-profile hacking incidents, including one in which OpenAI software broke out of a secure testing environment during a trial and independently accessed systems belonging to AI company Hugging Face.
Major US AI companies including OpenAI and Anthropic have since said they were prepared to slow the development of the technology, something critics had long called for.
