Azernews.Az

Ulviyya Poladova

Read more

Australian Prime Minister Anthony Albanese described as "unacceptable" an incident in which an OpenAI AI model allegedly bypassed security safeguards and gained unauthorized access to an Australian government website.

According to Albanese, in June, the AI tool attempted to access a health statistics portal. After being denied access, it allegedly bypassed the restrictions and entered a section containing confidential files.

"Today, I spoke with OpenAI CEO Sam Altman to express Australia’s deep concern over this incident. I also expressed disappointment that the company took too long to inform the government about what had happened," Albanese noted.

In addition to Medicare, the government is aware of three other systems that may have been affected. These include resources belonging to the Australian Institute of Health and Welfare, the NSW Bureau of Crime Statistics and Research, and the Victorian Department of Health.

In late July, The Guardian reported that an OpenAI AI agent involved in a series of attacks on the AI development platform Hugging Face had also accessed the system of OpenAI customer Modal Labs. According to Modal Labs, the agent exploited vulnerable code written by the company's customer and hosted on its platform. The agent reportedly escaped its isolated testing environment, or "sandbox," and breached another sandbox hosted within the infrastructure of a different provider.

Image: Dado Ruvic / Reuters