Newsletters
Australia's Medicare statistics database, which holds both public and non-public files, was breached by an artificial intelligence agent built by OpenAI, marking the first known case of AI hacking a government network, Prime Minister Anthony Albanese told reporters on the margins of the United Nations General Assembly in New York, following his discussion with OpenAI CEO Sam Altman.
The agent was also able to write files into the database, he said. The disclosure marks the latest incident involving AI agents behaving unpredictably and taking actions beyond what their human operators intended. It comes amid growing calls from leading figures in the AI industry, including Altman, for a slower pace of AI development.
Albanese said he conveyed Australia's "extreme concern" to Altman during a phone call. He also said Canberra was not aware of any previous case in which an AI system had breached a government system.
The Australian government has initiated an urgent review of its approach to artificial intelligence.
According to a government statement, the review will assess whether current laws and governance frameworks are "fit for purpose" when it comes to preparing for and responding to cyber incidents involving AI.
The Department of the Prime Minister and Cabinet will lead the review, which will also examine how Australia shares information with AI companies and other Commonwealth nations.
The government said the findings will contribute to its wider efforts to develop an effective framework for AI governance.
Andrew Charlton, the assistant minister for science, technology and the digital economy, told ABC that the country will impose "guardrails" on AI. "I think what we're observing at the moment is that a number of the AI models … inside their companies are not safe and they have a lot more work to do to make them safe before they release them to the public," he said.
Charlton said that the incident "also provides the justification … for our determination to make sure that we have Australian AI, that we shape technology, we have sovereign capability in Australia, and we aren't entirely at the mercy of foreign AI companies."
The move comes just a week after OpenAI released reports outlining six incidents involving unexpected or potentially concerning behavior by its AI models. The company also announced plans to monitor and publicly disclose similar incidents in the future.
The reports covered incidents identified between April and August. However, they do not appear to include the latest incident, which OpenAI said it became aware of in June.
Set as preferred source
Set as preferred source
1 min read
1 min read
