Marles calls unauthorized AI agent access ‘very serious’
Australian Deputy Prime Minister Richard Marles said Friday that OpenAI did not disclose to him at a meeting earlier this month that one of its artificial-intelligence agents had breached Australian government websites, after the company last week revealed the agent gained unauthorized access to the systems. Marles declined to speculate on whether Altman personally knew of the breach when the two met.
Marles told News 24 that the unauthorized access by an OpenAI agent is serious regardless of the data involved. “The fact that we’ve got an artificial intelligence agent gaining unauthorised access to an Australian government website, that of itself is very serious,” Marles said. “There’s no hiding that.”
The deputy prime minister said the information the agent obtained was “minor in its nature” and had already been made public by the government, adding that it did not include anyone’s personal data. But he said the seriousness lay in the breach itself. “In a sense, the seriousness doesn’t lie in the information that was obtained. It was the fact that we’ve got an unintended action of an AI agent gaining unauthorised access into an Australian government website,” Marles said.
Marles characterized the incident as a first for Australia. “This is the first time this has happened in the context of Australia, and so we’re not shying away from that,” he said.
OpenAI last week revealed that one of its AI agents had gained unauthorized access to Australian government websites.