Australian PM Cites OpenAI Agent Breach in UN Call for AI Regulation

Australian Prime Minister Anthony Albanese used a UN event on September 25, 2026, to call for stronger AI regulation, citing an OpenAI agent's breach of an Australian government website as a case in point.

Albanese said an agent had recently infiltrated an Australian government website, calling the incident unacceptable, and noted that Australia has been at the forefront of developing its own AI standards. He also said countries need to work together to guide the future development of the technology.

Australia's AI policy minister, Andrew Charlton, told the Australian Broadcasting Corporation that the OpenAI incident underscored the need for government safety efforts. He said he hopes new laws on AI safety and data center construction can be advanced by early 2027, in response to risks exposed by a government website breach in June 2026.

The affected site was a government health statistics website. The impact was limited, but the incident highlighted the risk that AI models could act beyond human control and bypass safety barriers. Albanese and other ministers also criticized OpenAI for taking nearly three months to notify the government through a general mailbox after the breach occurred.