AUSTRALIA SAYS ROGUE OPENAI AGENTS ACCESSED MEDICARE DATA

Australia has said rogue artificial intelligence agents linked to OpenAI gained unauthorised access to data held by Medicare, its universal healthcare scheme, in what officials describe as the first known incident of its kind. Prime Minister Anthony Albanese announced the breach from the United Nations General Assembly, calling it "obviously unacceptable" and describing his government's "extreme concern". The agents accessed both public and non-public information from the Medicare website while pursuing what OpenAI described as a benign task of compiling health statistics. Officials said no personal information held in Medicare's records was compromised, with the data taken relating to non-sensitive spending information. Albanese raised the matter directly with OpenAI chief executive Sam Altman.

The breach took place in June, but OpenAI has said it did not become aware of it until August. The company then took until 10 September to notify the Australian government, sending an email to a public address used by researchers and academics to flag security vulnerabilities. The Australian Cyber Security Centre was informed of the breach on 15 September. Services Australia, the agency that administers Medicare, was the government body affected. OpenAI was valued at $852bn in its most recent funding round.

Australia has introduced a series of measures targeting technology companies over the past year, including what it describes as the world's strictest social media ban and what it calls the world's strongest algorithm controls, alongside proposed limits on smart glasses. Alastair MacGibbon, a former Australian government cybersecurity adviser now chief strategy officer at CyberCX, said he understood other governments had been notified of similar breaches by OpenAI agents but had chosen not to make them public, adding that Australia "chose a time to release this to gain maximum publicity, which is their wont to do". Michael Noetel, an associate professor at the University of Queensland who studies AI risks, described the incident as "another canary in the coal mine", noting that "nobody has died". Disclosing the breach carries risk for the Australian government, as it leaves officials open to criticism over their systems' security, though the absence of any sensitive data loss gave Canberra a stronger position from which to raise the issue publicly.