Australia Says OpenAI Agent Breached Government Website

SYDNEY, Sept 24 (NNN-Kyodo) -- The Australian government revealed Thursday that an artificial intelligence agent developed by OpenAI infiltrated a government public health insurance statistics portal in June, Kyodo News reported.

Deputy Prime Minister Richard Marles told a press conference in Sydney that while the impact of the breach was minor, the unauthorised access by an AI agent was "a very serious incident" and the government is "extremely concerned."

According to the government, the AI agent accessed infrastructure behind the public-facing statistics portal which hosts non-sensitive data and statistics relating to the government's public health insurance scheme. No patient information is believed to have been accessed at this stage.

While the breach occurred in June, OpenAI only became aware of the matter in August during a review of "misaligned" model activity, and notified the Australian government agency responsible for the scheme by email this month.

During the incident, the AI agent interacted with four public Australian government websites, but only accessed publicly available information from the other three, according to Marles.

Prime Minister Anthony Albanese, who is in New York for the United Nations General Assembly, told a press conference that he held a phone call with OpenAI CEO Sam Altman Wednesday, New York time, to express Australia's "extreme concern" over the incident, and his disappointment that it took the company so long to inform the government.

A task force has been established to investigate the incident and the government's overall preparedness to respond to cyber incidents involving AI.

--NNN-KYODO