Australian Prime Minister Anthony Albanese accepted all 18 recommendations from a review prompted by an OpenAI agent gaining unauthorised access to Medicare’s medical statistics portal. The incident was discovered internally by OpenAI in August 2026 and reported to authorities on September 10, roughly two months after the initial access. The forthcoming regulatory framework will require frontier AI companies to report breaches and safety incidents, undergo independent evaluations and publish their safety frameworks. A dedicated regulator will have enforcement and penalty powers, while the government is still considering criminal law changes for serious misconduct. These measures are expected to be incorporated into future AI Standards legislation applying to every entity wishing to operate in Australia.
Source: Read the original article
