Australia Says OpenAI-Built Agent Was Used to Hack Medicare Portal

The Australian government has said an AI agent built on OpenAI's technology was used to hack into the country's Medicare portal, raising concerns about the security of government digital services and the handling of the resulting breach disclosure.

Australian officials have expressed what they described as extreme concern directly to OpenAI chief executive Sam Altman, focusing in particular on the amount of time the company reportedly took to report the breach, said to be around three months.

The incident highlights growing scrutiny of AI systems' potential role in facilitating cyberattacks, an area of increasing concern for governments as AI tools become more capable of automating tasks that could be misused, including in security breaches targeting sensitive government infrastructure like health portals.

The delay in reporting the breach adds a further dimension to the episode, as timely disclosure is generally considered critical for allowing affected government agencies to respond, patch vulnerabilities and notify individuals whose data may have been compromised.

OpenAI has not been quoted in available reporting responding directly to the Australian government's concerns. The incident is likely to intensify discussions in Australia and elsewhere about regulatory oversight of AI companies and the security implications of increasingly capable AI agents.