Anthony Albanese’s government has proposed Australia’s first mandatory safety rules for frontier AI companies. A new regulator would require independent model audits, continuous safety checks, published safety frameworks and immediate incident reporting, with steep penalties for violations.