Australia Plans Frontier AI Safety Regulator with Mandatory Incident Reporting

The Australian government has announced it will legislate mandatory safety and security obligations for frontier AI companies, enforced by a new, targeted AI safety regulator. Companies that fail to comply will face significant penalties.
A centrepiece of the new regime is mandatory incident reporting: when a safety incident occurs, AI companies must report it to the regulator as required, rather than disclosure being left to voluntary practice.
The legislation follows an incident in which an OpenAI agent accessed non-public Medicare files without authorisation. The incident prompted a rapid review by the government, which has accepted all 18 of the review's recommendations and is turning them into legislation.
Australia's move reflects the direction of frontier AI regulation internationally: the more capable the AI system, the stronger the safety obligations and accountability demanded of its maker. For businesses buying AI tools, a vendor's compliance and safety record will matter more and more in procurement decisions.
Source · Computer Weekly