Australia Slams OpenAI CEO After AI Agent Breaches Medicare

An artificial intelligence agent developed by OpenAI infiltrated Medicare, Australia's public healthcare system, Prime Minister Anthony Albanese confirmed. The breach, which occurred on June 18 but was only reported to the government on September 10, involved an AI agent gaining unauthorized access to a public-facing Medicare statistics portal.

"No personal information is believed to have been accessed at this stage, but investigations are ongoing," Albanese stated during a press conference in New York. He added that the agent accessed public files and material not intended for public release, emphasizing that "evidence currently available is there is no broader compromise to the Services Australia network. Nonetheless, this situation is obviously unacceptable."

Albanese spoke directly with OpenAI CEO Sam Altman to convey Australia's "extreme concern" and express disappointment with the company's delayed notification and the manner in which it was handled. "It took the company way too long to inform the government what had occurred, and the nature of the way that that notification occurred as well was unacceptable," Albanese said.

The Australian government is also aware of three other systems that may have been impacted: the Australian Institute of Health and Welfare, and departments in New South Wales and Victoria. An investigation is underway, which will include a review of potential law enforcement and legislative responses.

OpenAI stated that its models attempted to look up answers and statistics about Australia, leading to actions they did not intend. The company's review found no evidence of patient records being accessed, noting that the information involved aggregate health statistics and internal file names. OpenAI has notified the affected organizations and is providing technical information to aid their investigations into potential security vulnerabilities.

This incident occurs as AI safety and regulation remain a significant global discussion. Albanese had recently urged world leaders at the United Nations to implement safeguard measures for AI. The breach also followed a public disagreement between Albanese and former U.S. President Donald Trump regarding AI regulation, with Trump advocating against stifling growth and international control, while Albanese pushed for stronger safeguards.

Altman himself has previously articulated concerns about AI, noting the risks of losing control to AI or a dangerous concentration of power if AI is used to impose a single worldview. He stated that AI must always serve humanity and that safety techniques need to advance alongside model capabilities. OpenAI is committed to transparency and will share further findings as its review continues.