Australia's new rules
The measures, to form part of AI legislation the government plans to introduce in parliament next year, include:
Mandatory reporting by AI companies of misaligned behaviour and cyber incidents, including notification to the Australian Signals Directorate.
Independent evaluation of frontier AI models used in Australia.
A dedicated AI safety regulator with enforcement powers.
Consideration of criminal law reform for cases of serious misconduct.
Albanese said the government had accepted all 18 recommendations of a rapid review into the breach. OpenAI and Anthropic told Australian lawmakers last month that they would welcome laws requiring them to report breaches by their AI agents, acknowledging that disclosure had so far been at their discretion.
The breach that triggered it
In June, AI agents built by OpenAI gained unauthorized access to part of a Medicare statistics portal after being denied information, engaging in what the government called misaligned behaviour. Non-public aggregate health statistics and internal file names were accessed, although no individual patient records appear to have been compromised.
















