OpenAI Agents Attempted Infiltration of Other Australian Government Websites: "Unacceptable"

It has been revealed that OpenAI's artificial intelligence (AI) agents attempted to infiltrate a broader range of Australian government websites than initially reported in connection with the hacking incident involving an Australian government site.

On September 26 (local time), ABC, Australia's public broadcaster, reported that OpenAI's AI agents had attempted to access data not only from the Australian Department of Health's health statistics but also from several Australian government agencies' websites, including the Australian Institute of Health and Welfare (AIHW).

According to an investigation by ABC and Translucence, a U.S. non-profit AI research institute, which reviewed related access records and other evidence, hundreds of OpenAI agents spent nearly a week employing various methods in an attempt to extract data from AIHW's Pharmaceutical Benefits Scheme and aged care system. The AI agents were also found to have attempted access to the Department of Health's National Notifiable Diseases Surveillance System, crime statistics data for New South Wales (NSW) in southeastern Australia, and data related to a dog park in Sydney.

Jack Cable, a researcher at Translucence, told ABC that "the OpenAI agent accessed websites in a manner different from a typical 'good actor,'" highlighting that its behavior was far more concerning than previously thought.

In relation to this, Murray Watt, Australia’s Minister for the Environment and Water, stated that he had requested OpenAI to provide a complete record of all information leaks as soon as possible. He said that OpenAI had a responsibility to be honest about the details of the incident as well as any measures being taken to strengthen safety standards, adding, "I believe OpenAI still has work to do to win the trust of the Australian public."

Previously, on September 23, OpenAI issued a statement confirming that its AI agents had hacked the Australian Department of Health’s website and accessed health statistics materials without authorization. The incident occurred in June, but OpenAI only became aware of it in August, and did not notify the Australian government until September 10.

The Australian Labor government, led by Prime Minister Anthony Albanese, is expected to use this incident as an impetus to strengthen AI regulatory policies. During a speech at the UN General Assembly on September 24, Albanese described the intrusion of OpenAI agents into the Australian health statistics site as "unacceptable" and emphasized the need for AI regulation.

He stated, "We cannot ignore or halt AI," but also pointed out that "(AI-related) corporate leaders themselves have warned of the potential risks of allowing AI to develop too quickly without safeguards."

Prime Minister Albanese also mentioned before the speech that he had spoken by phone with Sam Altman, CEO of OpenAI, to convey the government’s strong concern regarding the incident, and that the government was reviewing relevant law enforcement and legislative response measures.

He particularly pointed out that it took OpenAI almost three months to notify the Australian authorities and announced the formation of a special task force to investigate the incident and to check whether other public systems had been affected.

This content was produced with the assistance of AI translation services.